Đại học Kinh tế Quốc dân – National Economics UniversityMICRO4ASIA
🏆 Arena
Sign inSign up
MICRO4ASIAThe six micro-credentials

Courses are delivered on the Limio learning platform. Limio

← Data Cleaning and Preprocessing with Python

You are viewing a public preview

Sign in to save your progress, do assignments, join discussions and get feedback.

Sign in
Start here

Course Syllabus

Print / Save as PDF

Course syllabus

This page restates the official course syllabus prepared by the instructor (draft of 7 November 2026). The course code and the alignment with programme learning outcomes (PLOs) are still to be completed by the responsible unit, so they are not shown here.

General information

ItemDetail
Course titleData Cleaning and Preprocessing with Python
ProgrammeMICRO4ASIA micro-credential
Credits1
Class time15 periods, organised as 5 sessions of 3 periods
Allocation7.5 periods of theory and instruction; 7.5 periods of practice, discussion and assessment
PrerequisitePython Programming for Data Science
InstructorDr Tran Duc Minh — minhdt@neu.edu.vn
FacultyFaculty of Data Science and Artificial Intelligence, College of Technology, National Economics University

Course description

This course equips you with foundational techniques for exploring, cleaning, preprocessing and interpreting data with Python. You use Pandas, NumPy, Matplotlib and Seaborn to examine tabular data, apply descriptive statistics and visualisation, address data quality issues and prepare data for reporting. Through a sales case study, you develop a rerunnable data processing workflow, validate the results and formulate evidence-based observations that support problem-solving in organisations.

Learning resources and software

  • Core materials: lecture videos and slides, instructional notebooks, a data dictionary, practice datasets and exercises for the five course topics.
  • Reference books: Wes McKinney (2022), Python for Data Analysis, 3rd edition, O'Reilly Media — Pandas fundamentals, data loading, cleaning and preparation, grouped aggregation, visualisation. Jake VanderPlas (2016), Python Data Science Handbook, O'Reilly Media — NumPy, Pandas and visualisation with Matplotlib.
  • Software: Python 3 and Google Colab with Pandas, NumPy, Matplotlib and Seaborn; openpyxl when you read or write Excel files.
  • Practice data: a sales transaction table with transaction ID, customer ID, transaction date, region, product category, quantity, unit price and discount rate. It deliberately contains missing values, duplicates, incorrect data types, inconsistent labels and unusual values. Each row represents one transaction for one product category; if real data are recorded at line-item level, the business key must be redefined.

Course goals

GoalDescription
G1Understand data quality and use Python to load and examine tabular data.
G2Apply descriptive statistics and visualisation to explore data.
G3Select and implement data cleaning techniques based on business context.
G4Perform preprocessing and create variables appropriate to the analytical objectives.
G5Develop a rerunnable workflow, validate outputs and report findings.

Course learning outcomes

CLOLearning outcomeBloom level
CLO1.1Explain the role of data quality and distinguish completeness, validity, consistency, uniqueness and accuracy.2 – Understand
CLO1.2Read CSV/Excel files, examine data structure and profile data quality using Pandas.3 – Apply
CLO2.1Calculate and interpret the mean, median, standard deviation, quantiles and IQR as appropriate to the data.3 – Apply
CLO2.2Analyse distributions, group differences and relationships using charts, summary tables and correlation.4 – Analyse
CLO3.1Handle missing values, duplicate records and inconsistent labels using explicit rules.3 – Apply
CLO3.2Convert data types, identify business rule violations and analyse outliers to propose appropriate treatments.4 – Analyse
CLO4.1Apply Min–Max scaling, Z-score standardisation and categorical encoding, and explain the purpose of each transformation.3 – Apply
CLO4.2Create revenue variables, time variables, value bands and summary tables for reporting.3 – Apply
CLO5.1Organise rerunnable notebooks, validate outputs and document changes before and after processing.3 – Apply
CLO5.2Interpret findings, evaluate data limitations and defend processing choices in a report.5 – Evaluate

How you are assessed

MethodDescriptionWhenCLOsWeight
Class participationAttendance, contributions to discussions and practical activities, and following class rules.Throughout the courseCLO1.1–CLO5.210%
Individual testsTwo computer-based tests in class, 20% each. Test 1: data loading, quality profiling, descriptive statistics and EDA. Test 2: missing values, duplicates, inconsistent labels, invalid types and outliers, with basic transformations and variable creation. You implement solutions in Python and briefly justify your choices.Test 1: Session 2, Period 6. Test 2: Session 4, Period 12.Test 1: CLO1.1–CLO2.2. Test 2: CLO3.1–CLO4.2.40%
Final examA 90-minute computer-based exam: explore, clean, preprocess and interpret a supplied dataset in Python, and submit a rerunnable notebook with quality checks, justified processing decisions, summary statistics and visualisations, and a concise interpretation. Eligibility: more than 80% attendance and both individual tests completed.University examination scheduleCLO1.1–CLO5.250%

Assessment indicators: correctness of Python operations, appropriateness of methods, accuracy of results, output validation, rerunnability and clarity of evidence-based explanations.

Teaching plan

SessionPeriods and contentCLOsPractice and evidence
1P1 Data quality and the workflow from raw data to reports. P2 DataFrame and Series; reading CSV/Excel; head, info, dtypes, shape; filtering and selecting. P3 Quality profiling with isna, duplicated, nunique, value_counts; business rules.CLO1.1, CLO1.2Exploration notebook and error table — short assignment 1
2P4 Mean, median, variance, standard deviation, quantiles, IQR. P5 Histograms, boxplots, bar charts and scatter plots. P6 groupby, agg, pivot_table; correlation and its limits.CLO2.1, CLO2.2Preliminary EDA report — short assignment 2; Test 1
3P7 Standardising missing markers; keeping, removing or imputing. P8 Full-row and key-based duplicates; standardising labels. P9 to_numeric, to_datetime; validating values and conversion errors.CLO3.1, CLO3.2Cleaned data and processing log — short assignment 3
4P10 IQR-based outliers versus data errors. P11 Min–Max, Z-score, one-hot and ordinal encoding. P12 Revenue, month/quarter and value bands; aggregating for reports.CLO3.2, CLO4.1, CLO4.2Preprocessing notebook and summary tables — short assignment 4; Test 2
5P13 Notebook organisation, functions, assert, logs and export. P14 Reporting findings, evidence and limitations. P15 Review, feedback and discussion.CLO5.1, CLO5.2Notebook, cleaned data, quality table and 1–2 page report — integrated assignment

Revenue in this course is defined as quantity × unit price × (1 − discount rate), with the discount rate in [0, 1].

Course requirements

  • Assignments and tests: four short practical assignments, an integrated practical assignment and individual tests. Deadlines, submission formats and permitted materials are announced in advance. Tests are completed independently on a computer with authorised materials only.
  • Participation: arrive on time with a working Python environment, read the materials and download the data before class, and take part in practice and discussion.
  • Quality and integrity: notebooks must rerun from the raw data and explain key decisions. Preserve the original files and document removed rows, modified values and unresolved issues. Do not alter data to reach a desired conclusion or assume that every outlier is an error. Acknowledge sources of code and materials and the extent of any AI assistance; you must be able to explain your code.
  • Asking for help: when you report an error, include a minimal code example, the error message and the result you expected.
  • Data and copyright: respect the copyright of lecture materials, notebooks and datasets. Use only simulated or authorised data, and remove personally identifiable information before sharing or uploading data.