pandas

Load, clean and analyze data with Python and pandas in a working notebook, from first steps to expert work.

Start lesson 1 Practice interview Take the assessmentPractice the first unit of every course. No account needed.

85lessons

526checked exercises

15projects, cases and exams

Prepares you for Take-home tasks

Projects, case studies and mock exams

Project · IntermediateCombine three store exports

Turn three differently formatted sales exports into one clean orders table and a product-by-month revenue report.

10 steps · 35 min Case study · IntermediateRegional targets

Compare monthly revenue with targets by region, handle an incomplete month and a month with no sales, and decide which region needs attention.

8 steps · 25 min Project · AdvancedA monthly KPI report

Build a month-by-month KPI table from orders, product costs and targets, leave out the incomplete month, and chart revenue against target.

10 steps · 35 min Case study · AdvancedThe spring promotion

Measure a discount promotion at one store against the weeks before it and against a store that had no promotion, and decide whether it paid off.

7 steps · 30 min Case study · AdvancedLate deliveries

Measure late deliveries by carrier, warehouse and dispatch day from UTC timestamps, separate two causes, and say what to fix first.

10 steps · 30 min Project · ExpertCustomer retention by signup month

Clean the membership orders, place every member in a signup-month cohort, and build a retention table that counts members who never ordered and leaves future months empty.

11 steps · 40 min Project · ExpertA repeatable cleaning pipeline

Turn a weekly supplier feed into clean data with small functions chained by pipe, report every problem in a checks table, and run the same pipeline on next week's file.

8 steps · 40 min Project · ExpertA web funnel report

Split a web event log into sessions with a 30-minute gap rule, mark which steps each session reached, and report view, cart and purchase rates by device.

9 steps · 35 min Case study · ExpertWhich products to reprice

Attach the price in effect to each week of sales, compare units at each price level within each product, and find which category loses the most volume when prices rise.

6 steps · 30 min Case study · ExpertWhat to reorder this week

Work out average daily demand from a sales log that skips days with no sales, compare days of stock with each supplier's lead time, and produce this week's reorder list.

6 steps · 25 min Mock exam · ExpertBike-share analysis

A take-home exercise as hiring teams set it: a bike-rental dataset, a brief with ten questions, and a 90-minute limit.

10 questions · 90 min Mock exam · ExpertTimed pandas assessment

Twelve independent questions of the kind online pandas assessments ask, across four datasets, with a 45-minute limit.

12 questions · 45 min

Practice pandas in a notebook that runs real Python in your browser, on realistic datasets from a homeware shop, a bike-rental company and a cafe. Every exercise opens with the data already loaded. You write the code, run the cell with Shift + Enter, and find out at once whether the result is correct, with a specific pointer when it isn't.

The course runs from Beginner to Expert: load and inspect a file, filter and sort, change and summarize columns, then clean messy exports, combine tables, reshape, work with time series, chart, and write fast, checked code. Each unit ends with a practice set on fresh data, and projects, case studies and interview sets put the skills together the way analysts use them at work and in hiring tests.

Beginner

First look at your data

Load a file, see what is in it, and pick out columns.

5 lessons

1Read a CSV file and look at itLoad a CSV file into a DataFrame with pd.read_csv, then check it with head, tail and shape.4 exercises · 6 minFree

2Inspect columns and typesCheck column types with dtypes and info, summarize numbers with describe, and count values with value_counts and nunique.5 exercises · 8 minFree

3Select columnsPick one column as a Series or several as a DataFrame, and drop the columns you don't need.5 exercises · 8 minFree

4Series, index and building a DataFrameSee how a column is a Series with an index, build small tables from dictionaries, and watch index alignment at work.5 exercises · 9 minFree

Practice setFirst lookMixed practice on a new dataset: size, types, counts, averages and column selection for a bike-rental shop.Free

Filter and sort

Keep the rows you need with masks, conditions and query, and put them in order.

6 lessons

6Filter rows with a conditionKeep only the rows that match a condition, like orders over 300 or orders from one region.5 exercises · 9 minWith a plan

7Combine conditions with & and |Filter on several conditions at once, match lists of values with isin, and exclude rows with ~.5 exercises · 10 minWith a plan

8Filter with queryWrite filters as readable strings with query, use Python variables with @, and handle column names that contain spaces.5 exercises · 9 minWith a plan

9Sort rowsOrder rows by one or more columns, largest or smallest first, and pick the top rows with nlargest.5 exercises · 9 minWith a plan

10Select with loc and ilocPick rows and columns by label with loc or by position with iloc, and change values safely with loc.5 exercises · 10 minWith a plan

Practice setFilter and sortMixed practice on bike rentals: masks, several conditions, comparing two columns, sorting and loc.With a plan

Change columns and rows

Add, rename, drop and convert columns, and remove duplicate rows.

5 lessons

12Add calculated columnsCreate new columns from existing ones with arithmetic, comparisons and np.where.5 exercises · 10 minWith a plan

13Rename and dropRename columns, tidy every header at once, and drop rows or columns by label, including ones that may not exist.5 exercises · 9 minWith a plan

14Convert column typesTurn text that looks like numbers into numbers, keep leading zeros, parse US dates and avoid the astype(bool) trap.5 exercises · 11 minWith a plan

15Find and remove duplicatesCount and remove duplicate rows, choose which copy to keep, and catch duplicates that differ only in capitals or spaces.5 exercises · 10 minWith a plan

Practice setChange columnsMixed practice on a cafe's orders: calculated columns, renames, conversions and duplicate checks.With a plan

Summarize

Group, aggregate, pivot and cross-tabulate.

5 lessons

17Summarize with groupbySplit rows into groups, such as regions, then add up, average or count each group.5 exercises · 10 minWith a plan

18Several summaries with aggCalculate several statistics per group in one step with agg, name the result columns, and flatten the result with reset_index.5 exercises · 11 minWith a plan

19Build pivot tablesTurn long data into a grid of totals, like a spreadsheet's PivotTable: rows, columns, values and an aggregation.5 exercises · 10 minWith a plan

20Count combinations with crosstabCount how often two categories occur together, turn counts into row or column percentages, and add totals.5 exercises · 9 minWith a plan

Practice setSummarizeMixed practice on cafe orders: totals by group, named results, pivot tables, crosstabs and groups with missing keys.With a plan

Intermediate

Clean messy data

Missing values, messy text, dates, recoding and bins.

6 lessons

22Find and fix missing valuesCount missing values, find the rows that have them, then drop or fill them on purpose.5 exercises · 10 minWith a plan

23Clean text with .str methodsTrim spaces, fix capitals, and search and replace text in whole columns with the .str accessor.5 exercises · 10 minWith a plan

24Dates with to_datetime and .dtTurn date text into real dates, pull out months, filter by date ranges and count the days between two dates.5 exercises · 11 minWith a plan

25Recode values with replace and mapSwap values with replace, look values up in a dictionary with map, and handle values the dictionary doesn't cover.5 exercises · 9 minWith a plan

26Put numbers into bins with cut and qcutGroup numbers into labeled bands with cut, choose which side of each border is included, and split into equal-size groups with qcut.5 exercises · 10 minWith a plan

Practice setClean messy dataClean a messy cafe export: count gaps, fix item and payment spellings, convert prices and mixed-format dates.With a plan

Combine tables

Join tables on keys, stack files, and work with the index.

4 lessons

28Join tables with mergeCombine two tables on a shared key column, choose the join type, and find rows that have no match.5 exercises · 12 minWith a plan

29Stack tables with concatStack files with the same columns into one table, renumber the index, tag where each row came from, and handle a column only one file has.5 exercises · 10 minWith a plan

30Index, set_index and joinMove a column into the index and back, look rows up by label, join tables on the index, and renumber rows after sorting.5 exercises · 10 minWith a plan

Practice setCombine tablesMixed practice joining bike rentals to stations: different key names, anti-joins, two joins with suffixes and IDs stored in different formats.With a plan

Functions and files

Run your own functions with map and apply, and read and write CSV, Excel, JSON and Parquet.

4 lessons

32apply, map and lambdaRun your own function on every value or every row with map and apply, and know when a built-in method is the better choice.5 exercises · 11 minWith a plan

33Read and write CSV filesWrite CSV files without the stray index column, and read files with semicolons, decimal commas, chosen columns and custom missing-value markers.5 exercises · 11 minWith a plan

34Excel, JSON and ParquetRead Excel sheets with title rows above the table, write Excel files, flatten nested JSON, and see why Parquet keeps your dates.5 exercises · 13 minWith a plan

Practice setFunctions and filesMixed practice on cafe and bike data: map and apply, CSV options, writing Excel and flattening JSON.With a plan

Projects and interview practice

Combine messy exports into a report, answer a targets question, and practice first-round interview questions.

3 lessons

ProjectCombine three store exportsTurn three differently formatted sales exports into one clean orders table and a product-by-month revenue report.With a plan

Case studyRegional targetsCompare monthly revenue with targets by region, handle an incomplete month and a month with no sales, and decide which region needs attention.With a plan

Interview setEasyEight first-round interview questions on an employee table: counts, averages, filters, the second-highest salary and ties.With a plan

Advanced

Reshape

Move between wide and long tables with melt, pivot, stack, unstack and explode.

5 lessons

39Wide to long with meltTurn a table with one column per month into one row per region and month, then sort and total it correctly.5 exercises · 10 minWith a plan

40Long to wide with pivotTurn long data wide with pivot, see why it fails on duplicate pairs, and know when pivot_table is the right tool.5 exercises · 10 minWith a plan

41stack and unstackMove an index level into the columns with unstack and back with stack, choose the level, and fill the gaps.5 exercises · 10 minWith a plan

42Lists in cells with explodeSplit cells that hold several values into lists, give each value its own row with explode, and handle blanks and stray separators.5 exercises · 10 minWith a plan

Practice setReshapeMixed reshaping practice on bike rentals, cafe orders, targets and product tags.With a plan

Group patterns

Group values on every row, keep whole groups, rank, run totals and find the top N per group.

5 lessons

44Group results on every row with transformPut a group total or average next to every row, work out shares of a group, compare rows with their group and fill gaps by group.5 exercises · 11 minWith a plan

45Whole groups with groupby.filterKeep only the groups that pass a test, such as customers with three or more orders, and know how it differs from filtering rows.5 exercises · 10 minWith a plan

46Rank, cumulative sums and countsRank values with the tie rule you need, rank within groups, and build running totals and order numbers.5 exercises · 11 minWith a plan

47Top N per groupFind the largest rows within each group, the top few per group, and handle ties at the cut-off.5 exercises · 11 minWith a plan

Practice setGroup patternsMixed practice on cafe orders and bike rentals: shares of group totals, group filters, ranks, running totals and top N per group.With a plan

Time series

Resample, rolling and expanding windows, growth rates and time zones.

6 lessons

49Resample by week and monthRoll daily figures up to weeks and months with resample, per group, and turn an irregular series into a regular daily one.5 exercises · 11 minWith a plan

50Rolling averagesSmooth daily figures with rolling averages, control the first rows with min_periods, and use calendar-day windows when days are missing.5 exercises · 11 minWith a plan

51Growth with shift and pct_changeCompare each period with the one before: change in units with diff, growth rates with pct_change, and any lag with shift.5 exercises · 11 minWith a plan

52Running totals and expanding windowsCompute figures "to date" with expanding windows, find record days with a running maximum, and smooth with an exponential average.5 exercises · 10 minWith a plan

53Time zones and timestampsRead UTC event times, convert them to local time across a daylight-saving change, and avoid comparing aware and naive times.5 exercises · 12 minWith a plan

Practice setTime seriesMixed time-series practice on two stores' daily figures: weekly totals, smoothing, growth, missing days and records.With a plan

Types, text and indexes

Ordered categories, MultiIndex selection and regular expressions.

4 lessons

55Ordered categoriesGive text values a meaningful order (S, M, L, XL), sort and compare by it, and show empty categories in counts.5 exercises · 11 minWith a plan

56MultiIndex selectionSelect from a two-level index by the outer level, the inner level or both, swap levels, and flatten two-level column names.5 exercises · 11 minWith a plan

57Text cleaning in depth with regexPull order numbers, emails and amounts out of free text with str.extract, match tags in any case, and avoid the special characters that trip regex up.5 exercises · 13 minWith a plan

Practice setTypes and textMixed practice: survey answers in scale order, two-level summaries, and values pulled out of ticket text.With a plan

Charts

Line, bar, histogram and box plots with df.plot, labeled so a reader can use them.

4 lessons

59Line and bar charts with df.plotDraw line and bar charts straight from a Series or DataFrame, sort the bars, and plot several lines at once.5 exercises · 11 minWith a plan

60Histograms, box plots and scatter plotsSee how values are spread with histograms and box plots, choose bins, and compare two columns with a scatter plot.5 exercises · 11 minWith a plan

61Label a chart for a reportGive charts a title and axis labels, work with the Axes object matplotlib returns, set the size and save the chart as an image.4 exercises · 10 minWith a plan

Practice setChartsMixed chart practice on sales, rentals and daily store figures, from picking the chart type to labeling it.With a plan

Projects and interview practice

Build a monthly KPI report, judge a promotion against a comparison store, find the causes of late deliveries, and practice second-round interview questions.

4 lessons

ProjectA monthly KPI reportBuild a month-by-month KPI table from orders, product costs and targets, leave out the incomplete month, and chart revenue against target.With a plan

Case studyThe spring promotionMeasure a discount promotion at one store against the weeks before it and against a store that had no promotion, and decide whether it paid off.With a plan

Case studyLate deliveriesMeasure late deliveries by carrier, warehouse and dispatch day from UTC timestamps, separate two causes, and say what to fix first.With a plan

Interview setMediumEight second-round questions: self-joins, group comparisons, ranks with ties, shares, growth, anti-joins and cohorts.With a plan

Expert

Merge edge cases

Joins that multiply rows, validate and indicator, and matching on the latest date with merge_asof.

4 lessons

67Duplicate keys and many-to-many joinsSpot joins that multiply rows because a key repeats, deduplicate or aggregate before joining, and join on every key that identifies a row.5 exercises · 12 minWith a plan

68Check joins with validate and indicatorMake a join fail loudly when keys aren't unique, find the record that breaks it, and see which rows matched with indicator.5 exercises · 11 minWith a plan

69Match to the latest price with merge_asofJoin each order to the price that was in effect on its date, per product, and control the direction and the gap allowed.6 exercises · 12 minWith a plan

Practice setMerge edge casesMixed practice on joins that go wrong: repeated keys, missing matches, validation and time-based joins.With a plan

Performance and style

Replace loops with vectorized code, shrink tables with dtypes, and write readable chains with assign and pipe.

4 lessons

71Vectorize instead of apply and iterrowsReplace row-by-row loops and apply with column arithmetic, np.where, np.select, map and groupby, which do the same work in compiled code.5 exercises · 13 minWith a plan

72Save memory with dtypesMeasure how much memory a table uses, find the columns that cost the most, and shrink them with categories, smaller number types and reading only what you need.6 exercises · 12 minWith a plan

73Method chaining with assign and pipeWrite an analysis as one readable chain: assign new columns, refer to them with lambda, filter with query, and slot your own functions in with pipe.5 exercises · 12 minWith a plan

Practice setPerformance and styleMixed practice: replace loops with vectorized code, shrink a table with dtypes, and write chains.With a plan

Data quality

Check keys, references and dates before you trust a file, and find outliers with the IQR rule, z-scores and group medians.

3 lessons

75Data quality checksCheck a dataset before you trust it: repeated rows, ids that point nowhere, dates in an impossible order, and a table that counts every problem.5 exercises · 13 minWith a plan

76Spot outliersFind unusual values with the IQR rule and z-scores, see why z-scores miss them in skewed data, cap values with clip, and catch a typing error within each group.5 exercises · 13 minWith a plan

Practice setData qualityMixed practice on a messy supplier feed and the membership tables: missing values, repeated keys, unknown references and a column report.With a plan

Projects, interviews and mock exams

Cohort retention, a reusable cleaning pipeline, a web funnel, pricing and reorder case studies, final-round interview questions and two timed mock exams.

8 lessons

ProjectCustomer retention by signup monthClean the membership orders, place every member in a signup-month cohort, and build a retention table that counts members who never ordered and leaves future months empty.With a plan

ProjectA repeatable cleaning pipelineTurn a weekly supplier feed into clean data with small functions chained by pipe, report every problem in a checks table, and run the same pipeline on next week's file.With a plan

ProjectA web funnel reportSplit a web event log into sessions with a 30-minute gap rule, mark which steps each session reached, and report view, cart and purchase rates by device.With a plan

Case studyWhich products to repriceAttach the price in effect to each week of sales, compare units at each price level within each product, and find which category loses the most volume when prices rise.With a plan

Case studyWhat to reorder this weekWork out average daily demand from a sales log that skips days with no sales, compare days of stock with each supplier's lead time, and produce this week's reorder list.With a plan

Interview setHardEight final-round questions: streaks, time to second order, rolling active users, products bought together, growth per group, top earners with ties, the top 20% of customers and lapsed customers.With a plan

Mock examBike-share analysisA take-home exercise as hiring teams set it: a bike-rental dataset, a brief with ten questions, and a 90-minute limit.With a plan

Mock examTimed pandas assessmentTwelve independent questions of the kind online pandas assessments ask, across four datasets, with a 45-minute limit.With a plan