Opening notes

Loading…

Notes PandasReal-World Pandas

Real-World PandasTopic 79 of 837 min read

Project: IPL Match Analysis

A mini-project inspired by notebook 06: load IPL match results, convert dates, and answer real questions — most wins, win %, toss decisions, wins by season, champions and top venues.

Notebook 06 of the course analyses the real IPL 2008–2020 match file (816 matches, 17 columns). Here we rebuild that analysis on a small 18-match sample (2018–2020) built inline, so every step runs in your browser. The columns match the real dataset, so the exact same code works on the full file.

text

         date team1 team2 toss_winner toss_decision winner
0  2018-04-07    MI   CSK         CSK         field    CSK
1  2018-04-08   KKR   RCB         KKR         field    KKR
2  2018-04-10   CSK   KKR         CSK         field    CSK
3  2018-04-21   RCB    DC         RCB         field    RCB
4  2018-05-05   SRH    DC          DC           bat    SRH

Definition

A mini-project applies the EDA workflow to one dataset to answer a fixed set of business questions. Here the dataset is one row per IPL match, with the teams (team1, team2), the toss (toss_winner, toss_decision = bat/field), the result (winner, result = won by runs/wickets, margin), the venue/city, and the player_of_match.

Our questions:

Which team won the most matches — and the highest percentage of matches played?
What do captains choose after winning the toss, and does winning the toss help?
How did each team do season by season, and who won each final?
Which venues hosted the most matches? What were the biggest wins?

Syntax — tools used in this project

python

pd.read_csv(io.StringIO(text))                 # load (a real file path works the same)
pd.to_datetime(s)  ·  s.dt.year  ·  s.dt.month_name()
s.value_counts(normalize=False)                 # counts (or proportions)
s1.add(s2, fill_value=0)                        # add two count Series safely
df.groupby(col)[col2].agg([...])                # per-group stats
df.pivot_table(index=, columns=, values=, aggfunc="count", fill_value=0)
pd.crosstab(row_series, col_series)             # frequency table
df.sort_values(...).groupby(...).last()         # last row per group (the final)

Function / parameter	Meaning	Default
`value_counts(normalize=)`	`True` → proportions that sum to 1 instead of counts	`False`
`Series.add(other, fill_value=)`	adds aligned on index; missing labels treated as `fill_value`	`None` (→ NaN)
`pivot_table(aggfunc=)`	how to combine values in each cell (`"count"`, `"sum"`, `"mean"`…)	`"mean"`
`pivot_table(fill_value=)`	value for empty cells	`None`
`pd.crosstab(index, columns)`	counts how often each combination occurs	—

Step 1 — load and fix the date type

python

import pandas as pd
import io

ipl = pd.read_csv(io.StringIO(IPL_CSV))     # the 18-match CSV text (see the examples)
print(ipl.shape)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year
print(ipl[["date", "season"]].dtypes)

text

(18, 11)
date      datetime64[us]
season             int32
dtype: object

How it works: read_csv reads dates as plain text (str). Converting with pd.to_datetime unlocks the .dt accessor, so .dt.year gives the season — exactly what notebook 06 does with ipl3['year'] = ipl3['date'].dt.year.

Step 2 — most wins and win percentage

python

print(ipl["winner"].value_counts())

played = ipl["team1"].value_counts().add(ipl["team2"].value_counts(), fill_value=0)
summary = pd.DataFrame({"played": played, "won": ipl["winner"].value_counts()}).fillna(0).astype(int)
summary["win_pct"] = (summary["won"] / summary["played"] * 100).round(1)
print(summary.sort_values("win_pct", ascending=False))

text

winner
CSK    5
MI     5
KKR    3
RCB    2
SRH    2
DC     1
Name: count, dtype: int64
     played  won  win_pct
CSK       6    5     83.3
MI        8    5     62.5
SRH       4    2     50.0
KKR       6    3     50.0
RCB       6    2     33.3
DC        6    1     16.7

How it works: a team can appear in either team1 or team2, so matches played = count in team1 + count in team2 (the notebook does the same with +). .add(..., fill_value=0) is safer than +: a team that only ever appears in one column would get NaN with +. CSK and MI both won 5, but CSK did it in 6 matches (83.3%) — raw counts can mislead; percentages are fairer.

Step 3 — the toss

python

print(ipl["toss_decision"].value_counts(normalize=True).round(2))
ipl["toss_win_match_win"] = ipl["toss_winner"] == ipl["winner"]
print(ipl["toss_win_match_win"].mean().round(2))
print(ipl.groupby("toss_decision")["toss_win_match_win"].agg(["count", "mean"]).round(2))

text

toss_decision
field    0.83
bat      0.17
Name: proportion, dtype: float64
0.5
               count  mean
toss_decision
bat                3  0.33
field             15  0.53

How it works: 83% of toss winners chose to field first. Comparing two columns gives a boolean Series, and the mean of booleans is the proportion of True — toss winners won exactly 50% of matches, so in this sample the toss gave no overall advantage. Choosing to field converted a bit better (53% vs 33%), but with only 3 "bat" matches that difference is not reliable — always check the count before trusting a percentage.

Step 4 — season by season, and the champions

python

print(ipl.pivot_table(index="winner", columns="season", values="venue",
                      aggfunc="count", fill_value=0))
champions = ipl.sort_values("date").groupby("season").last()[["winner", "venue"]]
print(champions)

text

season  2018  2019  2020
winner
CSK        3     1     1
DC         0     1     0
KKR        1     1     1
MI         0     2     3
RCB        1     0     1
SRH        1     1     0
       winner                        venue
season
2018      CSK             Wankhede Stadium
2019       MI         Rajiv Gandhi Stadium
2020       MI  Dubai International Stadium

How it works: pivot_table with aggfunc="count" counts rows for every (winner, season) pair — any non-null column works as values. fill_value=0 replaces "no wins" NaN with 0. For champions, the final is the last match of each season: sort by date, group by season, take .last() — the same idea as ipl4.groupby('year').last() in notebook 06. (Our sample is real-history friendly: CSK 2018, MI 2019 and 2020.)

Step 5 — venues and biggest wins

python

print(ipl["venue"].value_counts().head(3))
print(ipl.groupby("result")["margin"].agg(["count", "mean", "max"]).round(1))
print(ipl.loc[ipl["result"] == "runs"].nlargest(3, "margin")[["date", "winner", "team1", "team2", "margin"]])

text

venue
Dubai International Stadium    4
Wankhede Stadium               3
M Chinnaswamy Stadium          3
Name: count, dtype: int64
         count  mean  max
result
runs         7  31.3   59
wickets     11   5.6    9
         date winner team1 team2  margin
15 2020-10-24    KKR   KKR    DC      59
16 2020-11-05     MI    MI    DC      57
14 2020-09-23     MI   KKR    MI      49

How it works: a margin in runs and a margin in wickets are different units, so we group by result before averaging — mixing them would be meaningless. Teams batting first won 7 times (by 31.3 runs on average); chasing teams won 11 times.

Findings

Question	Answer (18-match sample)
Most wins	CSK and MI, 5 each — CSK has the best win % (83.3%)
Toss choice	83% chose to field
Does the toss help?	Toss winners won 50% — no clear advantage
Champions	CSK (2018), MI (2019, 2020)
Busiest venue	Dubai International Stadium (4 matches)
Biggest win	KKR by 59 runs vs DC (2020-10-24)

Note (real data): the full 2008–2020 file needs extra cleaning that notebook 06 shows: fillna({"method": "fare_match"}) before dropna(), merging renamed teams with replace({"Delhi Daredevils": "Delhi Capitals", ...}), and fixing cities like "Bengaluru" → "Bangalore".

Common mistakes

❌ Counting matches played from team1 only. ✅ Add team1 and team2 counts.
❌ Ranking by raw wins when teams played different numbers of matches. ✅ Compare win percentage.
❌ Averaging margin across runs and wickets. ✅ Group by result first.
❌ Using .last() without sorting by date. ✅ sort_values("date") first, so "last" really is the final.
❌ Forgetting renamed franchises (Delhi Daredevils → Delhi Capitals) in the real data. ✅ Unify names with replace before grouping.

Interview tip

Q: How do you test whether winning the toss helps win the match? A: Create a boolean column toss_winner == winner and take its mean() — that is the share of matches the toss winner won. Then groupby("toss_decision") (or by season/venue) to see where it differs, and always look at the group count before trusting small-sample differences.

Key points to remember

Convert date text with pd.to_datetime, then derive season with .dt.year — the first fix in any match dataset.
Matches played = team1 counts + team2 counts; use .add(other, fill_value=0) to avoid NaN.
Compare win percentage, not raw wins, when teams played different numbers of matches.
The mean of a boolean column is the proportion of True — e.g. how often the toss winner won.
pivot_table(index='winner', columns='season', aggfunc='count', fill_value=0) builds a wins-by-season table.
Sort by date then groupby('season').last() gives each season's final.
Group by result type before averaging margins — runs and wickets are different units.

Try it yourself

Load, convert dates, most wins and win %

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl[["date", "team1", "team2", "winner"]].head())
print(ipl["winner"].value_counts())

played = ipl["team1"].value_counts().add(ipl["team2"].value_counts(), fill_value=0)
summary = pd.DataFrame({"played": played, "won": ipl["winner"].value_counts()}).fillna(0).astype(int)
summary["win_pct"] = (summary["won"] / summary["played"] * 100).round(1)
print(summary.sort_values("win_pct", ascending=False))

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl[["date", "team1", "team2", "winner"]].head())
print(ipl["winner"].value_counts())

played = ipl["team1"].value_counts().add(ipl["team2"].value_counts(), fill_value=0)
summary = pd.DataFrame({"played": played, "won": ipl["winner"].value_counts()}).fillna(0).astype(int)
summary["win_pct"] = (summary["won"] / summary["played"] * 100).round(1)
print(summary.sort_values("win_pct", ascending=False))

Toss decisions and toss advantage

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl["toss_decision"].value_counts())
print(ipl["toss_decision"].value_counts(normalize=True).round(2))

ipl["toss_win_match_win"] = ipl["toss_winner"] == ipl["winner"]
print("toss winner won:", ipl["toss_win_match_win"].mean().round(2))
print(ipl.groupby("toss_decision")["toss_win_match_win"].agg(["count", "mean"]).round(2))
print(pd.crosstab(ipl["season"], ipl["toss_decision"]))

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl["toss_decision"].value_counts())
print(ipl["toss_decision"].value_counts(normalize=True).round(2))

ipl["toss_win_match_win"] = ipl["toss_winner"] == ipl["winner"]
print("toss winner won:", ipl["toss_win_match_win"].mean().round(2))
print(ipl.groupby("toss_decision")["toss_win_match_win"].agg(["count", "mean"]).round(2))
print(pd.crosstab(ipl["season"], ipl["toss_decision"]))

Wins by season, champions and venues

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl.pivot_table(index="winner", columns="season", values="venue",
                      aggfunc="count", fill_value=0))

champions = ipl.sort_values("date").groupby("season").last()[["winner", "venue"]]
print(champions)

print(ipl["venue"].value_counts().head(3))
print(ipl.groupby("result")["margin"].agg(["count", "mean", "max"]).round(1))
print(ipl["player_of_match"].value_counts().head(3))

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl.pivot_table(index="winner", columns="season", values="venue",
                      aggfunc="count", fill_value=0))

champions = ipl.sort_values("date").groupby("season").last()[["winner", "venue"]]
print(champions)

print(ipl["venue"].value_counts().head(3))
print(ipl.groupby("result")["margin"].agg(["count", "mean", "max"]).round(1))
print(ipl["player_of_match"].value_counts().head(3))

Practice: your own IPL question

# 🏏 Your turn: which month had the most matches? Which team won most often while CHASING (result == "wickets")? Edit and Run.
import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

ipl["month"] = ipl["date"].dt.month_name()
print(ipl["month"].value_counts())

chasing_wins = ipl.loc[ipl["result"] == "wickets", "winner"].value_counts()
print(chasing_wins)

# 🏏 Your turn: which month had the most matches? Which team won most often while CHASING (result == "wickets")? Edit and Run.
import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

ipl["month"] = ipl["date"].dt.month_name()
print(ipl["month"].value_counts())

chasing_wins = ipl.loc[ipl["result"] == "wickets", "winner"].value_counts()
print(chasing_wins)

Revision Flashcards

1 / 6

Tap a card to flip it and see the answer.

Opening notes

Notes PandasReal-World Pandas

Real-World PandasTopic 79 of 837 min read

Project: IPL Match Analysis

A mini-project inspired by notebook 06: load IPL match results, convert dates, and answer real questions — most wins, win %, toss decisions, wins by season, champions and top venues.

text

         date team1 team2 toss_winner toss_decision winner
0  2018-04-07    MI   CSK         CSK         field    CSK
1  2018-04-08   KKR   RCB         KKR         field    KKR
2  2018-04-10   CSK   KKR         CSK         field    CSK
3  2018-04-21   RCB    DC         RCB         field    RCB
4  2018-05-05   SRH    DC          DC           bat    SRH

Definition

A mini-project applies the EDA workflow to one dataset to answer a fixed set of business questions. Here the dataset is one row per IPL match, with the teams (team1, team2), the toss (toss_winner, toss_decision = bat/field), the result (winner, result = won by runs/wickets, margin), the venue/city, and the player_of_match.

Our questions:

Which team won the most matches — and the highest percentage of matches played?
What do captains choose after winning the toss, and does winning the toss help?
How did each team do season by season, and who won each final?
Which venues hosted the most matches? What were the biggest wins?

Syntax — tools used in this project

python

pd.read_csv(io.StringIO(text))                 # load (a real file path works the same)
pd.to_datetime(s)  ·  s.dt.year  ·  s.dt.month_name()
s.value_counts(normalize=False)                 # counts (or proportions)
s1.add(s2, fill_value=0)                        # add two count Series safely
df.groupby(col)[col2].agg([...])                # per-group stats
df.pivot_table(index=, columns=, values=, aggfunc="count", fill_value=0)
pd.crosstab(row_series, col_series)             # frequency table
df.sort_values(...).groupby(...).last()         # last row per group (the final)

Function / parameter	Meaning	Default
`value_counts(normalize=)`	`True` → proportions that sum to 1 instead of counts	`False`
`Series.add(other, fill_value=)`	adds aligned on index; missing labels treated as `fill_value`	`None` (→ NaN)
`pivot_table(aggfunc=)`	how to combine values in each cell (`"count"`, `"sum"`, `"mean"`…)	`"mean"`
`pivot_table(fill_value=)`	value for empty cells	`None`
`pd.crosstab(index, columns)`	counts how often each combination occurs	—

Step 1 — load and fix the date type

python

import pandas as pd
import io

ipl = pd.read_csv(io.StringIO(IPL_CSV))     # the 18-match CSV text (see the examples)
print(ipl.shape)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year
print(ipl[["date", "season"]].dtypes)

text

(18, 11)
date      datetime64[us]
season             int32
dtype: object

Step 2 — most wins and win percentage

python

print(ipl["winner"].value_counts())

played = ipl["team1"].value_counts().add(ipl["team2"].value_counts(), fill_value=0)
summary = pd.DataFrame({"played": played, "won": ipl["winner"].value_counts()}).fillna(0).astype(int)
summary["win_pct"] = (summary["won"] / summary["played"] * 100).round(1)
print(summary.sort_values("win_pct", ascending=False))

text

winner
CSK    5
MI     5
KKR    3
RCB    2
SRH    2
DC     1
Name: count, dtype: int64
     played  won  win_pct
CSK       6    5     83.3
MI        8    5     62.5
SRH       4    2     50.0
KKR       6    3     50.0
RCB       6    2     33.3
DC        6    1     16.7

Step 3 — the toss

python

print(ipl["toss_decision"].value_counts(normalize=True).round(2))
ipl["toss_win_match_win"] = ipl["toss_winner"] == ipl["winner"]
print(ipl["toss_win_match_win"].mean().round(2))
print(ipl.groupby("toss_decision")["toss_win_match_win"].agg(["count", "mean"]).round(2))

text

toss_decision
field    0.83
bat      0.17
Name: proportion, dtype: float64
0.5
               count  mean
toss_decision
bat                3  0.33
field             15  0.53

Step 4 — season by season, and the champions

python

print(ipl.pivot_table(index="winner", columns="season", values="venue",
                      aggfunc="count", fill_value=0))
champions = ipl.sort_values("date").groupby("season").last()[["winner", "venue"]]
print(champions)

text

season  2018  2019  2020
winner
CSK        3     1     1
DC         0     1     0
KKR        1     1     1
MI         0     2     3
RCB        1     0     1
SRH        1     1     0
       winner                        venue
season
2018      CSK             Wankhede Stadium
2019       MI         Rajiv Gandhi Stadium
2020       MI  Dubai International Stadium

Step 5 — venues and biggest wins

python

print(ipl["venue"].value_counts().head(3))
print(ipl.groupby("result")["margin"].agg(["count", "mean", "max"]).round(1))
print(ipl.loc[ipl["result"] == "runs"].nlargest(3, "margin")[["date", "winner", "team1", "team2", "margin"]])

text

venue
Dubai International Stadium    4
Wankhede Stadium               3
M Chinnaswamy Stadium          3
Name: count, dtype: int64
         count  mean  max
result
runs         7  31.3   59
wickets     11   5.6    9
         date winner team1 team2  margin
15 2020-10-24    KKR   KKR    DC      59
16 2020-11-05     MI    MI    DC      57
14 2020-09-23     MI   KKR    MI      49

Findings

Question	Answer (18-match sample)
Most wins	CSK and MI, 5 each — CSK has the best win % (83.3%)
Toss choice	83% chose to field
Does the toss help?	Toss winners won 50% — no clear advantage
Champions	CSK (2018), MI (2019, 2020)
Busiest venue	Dubai International Stadium (4 matches)
Biggest win	KKR by 59 runs vs DC (2020-10-24)

Note (real data): the full 2008–2020 file needs extra cleaning that notebook 06 shows: fillna({"method": "fare_match"}) before dropna(), merging renamed teams with replace({"Delhi Daredevils": "Delhi Capitals", ...}), and fixing cities like "Bengaluru" → "Bangalore".

Common mistakes

❌ Counting matches played from team1 only. ✅ Add team1 and team2 counts.
❌ Ranking by raw wins when teams played different numbers of matches. ✅ Compare win percentage.
❌ Averaging margin across runs and wickets. ✅ Group by result first.
❌ Using .last() without sorting by date. ✅ sort_values("date") first, so "last" really is the final.
❌ Forgetting renamed franchises (Delhi Daredevils → Delhi Capitals) in the real data. ✅ Unify names with replace before grouping.

Interview tip

Key points to remember

Convert date text with pd.to_datetime, then derive season with .dt.year — the first fix in any match dataset.
Matches played = team1 counts + team2 counts; use .add(other, fill_value=0) to avoid NaN.
Compare win percentage, not raw wins, when teams played different numbers of matches.
The mean of a boolean column is the proportion of True — e.g. how often the toss winner won.
pivot_table(index='winner', columns='season', aggfunc='count', fill_value=0) builds a wins-by-season table.
Sort by date then groupby('season').last() gives each season's final.
Group by result type before averaging margins — runs and wickets are different units.

Try it yourself

Load, convert dates, most wins and win %

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl[["date", "team1", "team2", "winner"]].head())
print(ipl["winner"].value_counts())

played = ipl["team1"].value_counts().add(ipl["team2"].value_counts(), fill_value=0)
summary = pd.DataFrame({"played": played, "won": ipl["winner"].value_counts()}).fillna(0).astype(int)
summary["win_pct"] = (summary["won"] / summary["played"] * 100).round(1)
print(summary.sort_values("win_pct", ascending=False))

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl[["date", "team1", "team2", "winner"]].head())
print(ipl["winner"].value_counts())

played = ipl["team1"].value_counts().add(ipl["team2"].value_counts(), fill_value=0)
summary = pd.DataFrame({"played": played, "won": ipl["winner"].value_counts()}).fillna(0).astype(int)
summary["win_pct"] = (summary["won"] / summary["played"] * 100).round(1)
print(summary.sort_values("win_pct", ascending=False))

Toss decisions and toss advantage

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl["toss_decision"].value_counts())
print(ipl["toss_decision"].value_counts(normalize=True).round(2))

ipl["toss_win_match_win"] = ipl["toss_winner"] == ipl["winner"]
print("toss winner won:", ipl["toss_win_match_win"].mean().round(2))
print(ipl.groupby("toss_decision")["toss_win_match_win"].agg(["count", "mean"]).round(2))
print(pd.crosstab(ipl["season"], ipl["toss_decision"]))

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl["toss_decision"].value_counts())
print(ipl["toss_decision"].value_counts(normalize=True).round(2))

ipl["toss_win_match_win"] = ipl["toss_winner"] == ipl["winner"]
print("toss winner won:", ipl["toss_win_match_win"].mean().round(2))
print(ipl.groupby("toss_decision")["toss_win_match_win"].agg(["count", "mean"]).round(2))
print(pd.crosstab(ipl["season"], ipl["toss_decision"]))

Wins by season, champions and venues

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl.pivot_table(index="winner", columns="season", values="venue",
                      aggfunc="count", fill_value=0))

champions = ipl.sort_values("date").groupby("season").last()[["winner", "venue"]]
print(champions)

print(ipl["venue"].value_counts().head(3))
print(ipl.groupby("result")["margin"].agg(["count", "mean", "max"]).round(1))
print(ipl["player_of_match"].value_counts().head(3))

import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

print(ipl.pivot_table(index="winner", columns="season", values="venue",
                      aggfunc="count", fill_value=0))

champions = ipl.sort_values("date").groupby("season").last()[["winner", "venue"]]
print(champions)

print(ipl["venue"].value_counts().head(3))
print(ipl.groupby("result")["margin"].agg(["count", "mean", "max"]).round(1))
print(ipl["player_of_match"].value_counts().head(3))

Practice: your own IPL question

# 🏏 Your turn: which month had the most matches? Which team won most often while CHASING (result == "wickets")? Edit and Run.
import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

ipl["month"] = ipl["date"].dt.month_name()
print(ipl["month"].value_counts())

chasing_wins = ipl.loc[ipl["result"] == "wickets", "winner"].value_counts()
print(chasing_wins)

# 🏏 Your turn: which month had the most matches? Which team won most often while CHASING (result == "wickets")? Edit and Run.
import pandas as pd
import io

csv = io.StringIO("""date,city,venue,team1,team2,toss_winner,toss_decision,winner,result,margin,player_of_match
2018-04-07,Mumbai,Wankhede Stadium,MI,CSK,CSK,field,CSK,wickets,1,DJ Bravo
2018-04-08,Kolkata,Eden Gardens,KKR,RCB,KKR,field,KKR,wickets,4,SP Narine
2018-04-10,Chennai,MA Chidambaram Stadium,CSK,KKR,CSK,field,CSK,wickets,5,SW Billings
2018-04-21,Bengaluru,M Chinnaswamy Stadium,RCB,DC,RCB,field,RCB,wickets,6,AB de Villiers
2018-05-05,Hyderabad,Rajiv Gandhi Stadium,SRH,DC,DC,bat,SRH,wickets,7,Rashid Khan
2018-05-27,Mumbai,Wankhede Stadium,SRH,CSK,CSK,field,CSK,wickets,8,SR Watson
2019-03-23,Chennai,MA Chidambaram Stadium,CSK,RCB,CSK,field,CSK,wickets,7,Harbhajan Singh
2019-03-24,Mumbai,Wankhede Stadium,MI,DC,MI,field,DC,runs,37,RR Pant
2019-03-28,Bengaluru,M Chinnaswamy Stadium,RCB,MI,RCB,field,MI,runs,6,JJ Bumrah
2019-04-05,Bengaluru,M Chinnaswamy Stadium,RCB,KKR,KKR,field,KKR,wickets,5,AD Russell
2019-04-29,Hyderabad,Rajiv Gandhi Stadium,SRH,KKR,KKR,field,SRH,wickets,9,DA Warner
2019-05-12,Hyderabad,Rajiv Gandhi Stadium,MI,CSK,MI,bat,MI,runs,1,JJ Bumrah
2020-09-19,Abu Dhabi,Sheikh Zayed Stadium,MI,CSK,CSK,field,CSK,wickets,5,AT Rayudu
2020-09-21,Dubai,Dubai International Stadium,SRH,RCB,SRH,field,RCB,runs,10,YS Chahal
2020-09-23,Abu Dhabi,Sheikh Zayed Stadium,KKR,MI,KKR,field,MI,runs,49,RG Sharma
2020-10-24,Dubai,Dubai International Stadium,KKR,DC,DC,field,KKR,runs,59,CV Varun
2020-11-05,Dubai,Dubai International Stadium,MI,DC,DC,field,MI,runs,57,JJ Bumrah
2020-11-10,Dubai,Dubai International Stadium,DC,MI,DC,bat,MI,wickets,5,TA Boult""")
ipl = pd.read_csv(csv)
ipl["date"] = pd.to_datetime(ipl["date"])
ipl["season"] = ipl["date"].dt.year

ipl["month"] = ipl["date"].dt.month_name()
print(ipl["month"].value_counts())

chasing_wins = ipl.loc[ipl["result"] == "wickets", "winner"].value_counts()
print(chasing_wins)

Revision Flashcards

1 / 6

Tap a card to flip it and see the answer.