DSC 101 Introduction To Data Science

Course Overview

Data science is the disciplined process of using data to understand situations, answer questions, support decisions and build useful predictive systems. It combines ideas from statistics, computing, mathematics and subject-matter knowledge. This course introduces the field at a beginner-friendly level and prepares students for later study in programming, databases, statistics, machine learning, data engineering and artificial intelligence.

The course emphasises responsible problem-solving. Students will learn that good data science is not merely the use of software; it begins with a clear question, suitable data, careful analysis and honest communication of limitations. Examples are drawn from education, health, agriculture, banking, telecommunications, public administration and other areas relevant to Nigeria and the wider world.

Learning Outcomes

• By the end of this course, students should be able to explain the meaning, scope and major stages of data science;
• By the end of this course, students should be able to distinguish common types, structures and sources of data;
• By the end of this course, students should be able to describe methods of data collection and recognise data-quality problems;
• By the end of this course, students should be able to apply basic ethical, privacy and security principles when working with data;
• By the end of this course, students should be able to use spreadsheets, databases and introductory programming tools for simple data tasks;
• By the end of this course, students should be able to clean, organise and summarise small datasets;
• By the end of this course, students should be able to interpret basic descriptive statistics and common data visualisations;
• By the end of this course, students should be able to explain foundational probability and statistical-inference ideas;
• By the end of this course, students should be able to describe the purpose of machine learning and distinguish major learning approaches;
• By the end of this course, students should be able to communicate data findings clearly and identify careers in the data profession.

Unit 1: Meaning, Scope and Evolution of Data Science

Data are recorded facts, measurements, observations or symbols that can be processed to produce information. Data science is the systematic study of how data are collected, managed, analysed and communicated in order to create knowledge or support action. A data scientist therefore does more than calculate numbers: the scientist frames questions, understands the context, selects methods, checks assumptions and explains results.

Modern data science grew from statistics, database management, scientific computing and artificial intelligence. Organisations now produce large quantities of data through mobile phones, websites, sensors, financial transactions, health systems and social platforms. The growing availability of computing power and storage has made it possible to analyse these records quickly, but large datasets do not automatically produce reliable conclusions.

The data science lifecycle commonly includes problem definition, data collection, data preparation, exploration, modelling, evaluation, communication and deployment. These stages are iterative. A weak result may require the analyst to obtain better data, revise the question or select a different method.

Key Takeaways

Check Your Understanding

Unit 2: Types, Structures and Sources of Data

Data may be qualitative or quantitative. Qualitative data describe categories or qualities, such as department, blood group or customer opinion. Quantitative data represent amounts or measurements. Quantitative data may be discrete, such as the number of students in a class, or continuous, such as height, temperature or travel time.

Measurement scales affect the analyses that are appropriate. Nominal data consist of labels without order. Ordinal data have a meaningful ranking. Interval data have equal differences but no true zero, while ratio data have equal intervals and a meaningful zero. A student should not compute an average from arbitrary category codes simply because they are written as numbers.

Structured data fit a predefined arrangement such as rows and columns in a table. Semi-structured data contain tags or keys, as in JSON or XML files. Unstructured data include free text, images, audio and video. Sources may be primary, when collected directly for a study, or secondary, when obtained from existing records, open-data portals, reports or previous research.

Key Takeaways

Check Your Understanding

Unit 3: Data Collection, Quality, Ethics and Privacy

Data may be collected through surveys, interviews, experiments, observation, sensors, transaction systems, application logs and administrative records. A census attempts to study an entire population, while a sample studies a smaller group. A good sample should represent the population sufficiently for the intended purpose. Convenience samples are easy to obtain but may produce biased conclusions.

Data quality is commonly assessed through accuracy, completeness, consistency, timeliness, validity and uniqueness. Missing values, duplicate records, impossible ages, inconsistent spellings and incorrect dates can distort analysis. Data provenance records where data came from, how they were created and what changes were made to them.

Ethical data practice requires a legitimate purpose, appropriate consent or other lawful basis, collection of only necessary data, protection against unauthorised access and respect for the rights of individuals. Sensitive data should not be exposed in classroom projects. Removing names may not be enough because combinations such as location, age and occupation can sometimes re-identify a person. Analysts must also consider fairness: biased data may produce decisions that disadvantage particular groups.

Key Takeaways

Check Your Understanding

Unit 4: Tools for Working with Data

Spreadsheets are useful for entering, organising, filtering, calculating and charting small or medium-sized datasets. Common functions include SUM, AVERAGE, COUNT, IF and lookup operations. However, a spreadsheet can become difficult to audit when formulas are copied incorrectly or many people edit separate versions.

Relational databases store data in linked tables and are queried with Structured Query Language (SQL). Databases are more suitable than spreadsheets for large, shared or frequently updated records. Programming languages such as Python and R support reproducible analysis and automation. Python libraries commonly used in later courses include pandas for tabular data, NumPy for numerical work and Matplotlib for visualisation.

A reproducible analysis can be repeated with the same steps and data to obtain the same result. Clear file names, version control, documented assumptions and scripts are therefore valuable. Students should avoid changing the only copy of raw data; instead, preserve the original and create a working copy.

# A simple Python example
scores = [62, 75, 48, 81, 69]
average = sum(scores) / len(scores)
print(round(average, 2))

Key Takeaways

Check Your Understanding

Unit 5: Data Cleaning and Preparation

Data cleaning is the process of identifying and correcting problems that could weaken analysis. Typical tasks include removing duplicates, standardising names and dates, correcting invalid values, handling missing observations and converting variables into suitable formats. Cleaning decisions should be recorded rather than hidden.

Missing data may be left as missing, removed, or replaced through a justified method called imputation. The correct choice depends on why the values are absent and how much information would be lost. Automatically replacing every missing value with zero is usually incorrect because zero is a real measurement, not a symbol for “unknown”.

Data transformation changes the form of data to make analysis easier. Examples include creating age groups, calculating totals, normalising numerical values and encoding categories. Outliers are unusually large or small observations. They may be errors, rare but genuine cases, or important signals; therefore, they should be investigated before removal.

Key Takeaways

Check Your Understanding

Unit 6: Descriptive Statistics and Data Visualisation

Descriptive statistics summarise the main features of a dataset. Measures of central tendency include the mean, median and mode. The mean uses every value but can be strongly affected by extreme observations. The median is the middle value after ordering the data and is often more suitable for skewed variables such as income. The mode is the most frequent value and can be used with categorical data.

Measures of spread describe variability. The range is the difference between the maximum and minimum. Variance and standard deviation measure how far values tend to lie from the mean. Percentiles and the interquartile range describe position and are less sensitive to extreme values. A summary should normally report both centre and spread.

Visualisation presents patterns in a form that can be quickly understood. Bar charts compare categories, histograms show distributions of numerical variables, line charts show change over time, scatter plots show relationships between two numerical variables, and box plots summarise distributions and possible outliers. Charts should have clear titles, labelled axes, suitable scales and honest proportions.

Key Takeaways

Check Your Understanding

Unit 7: Probability, Sampling and Statistical Inference

Probability measures uncertainty on a scale from 0 to 1. An impossible event has probability 0, while a certain event has probability 1. Conditional probability is the probability of an event given that another event has occurred. Independence means that knowing one event occurred does not change the probability of the other.

A population contains all units of interest, while a sample is the subset actually observed. Sampling variation means that different samples from the same population may produce different results. Random sampling helps reduce systematic selection bias, although it does not guarantee a perfect sample.

Statistical inference uses sample data to estimate population characteristics or test claims. A confidence interval gives a range of plausible values for a population parameter. A hypothesis test evaluates how compatible the observed data are with a stated claim. Statistical significance does not automatically mean practical importance, and association does not by itself prove causation.

Key Takeaways

Check Your Understanding

Unit 8: Introduction to Machine Learning

Machine learning is a branch of artificial intelligence in which algorithms learn patterns from data in order to make predictions, classifications or decisions. In supervised learning, examples include both input variables and a known target, such as past loan records labelled “repaid” or “defaulted”. Classification predicts categories, while regression predicts numerical values.

Unsupervised learning explores data without a known target. Clustering groups similar observations, while dimensionality reduction creates simpler representations of complex data. Reinforcement learning involves an agent that learns by receiving rewards or penalties from an environment.

A dataset is often divided into training and test portions. The model learns from the training data and is evaluated on unseen test data. Overfitting occurs when a model memorises the training data and performs poorly on new cases. Evaluation measures depend on the task and should reflect real consequences. Accuracy alone can be misleading when one class is rare.

Key Takeaways

Check Your Understanding

Unit 9: Databases, Big Data and Data Engineering

Data engineering is concerned with building reliable systems that collect, store, transform and deliver data for analysis. A data pipeline moves data from sources through processing stages to a destination such as a database, warehouse or dashboard. Data scientists depend on well-designed pipelines because models cannot compensate for unavailable or unreliable data.

Relational databases organise data into tables connected by keys. Data warehouses combine historical data for reporting and analysis, while data lakes store large amounts of raw or semi-processed data in many formats. Big data is often described through high volume, velocity and variety, together with concerns about veracity and value.

Cloud platforms provide scalable storage and computing resources over the Internet. They can reduce the need for local infrastructure, but organisations must manage cost, access control, backups, data location and regulatory obligations. Big data should not be collected merely because storage is available; usefulness, quality and lawful purpose remain essential.

Key Takeaways

Check Your Understanding

Unit 10: Communication, Nigerian Applications and Data Careers

The final product of data work is often a decision, not a model. Data storytelling combines evidence, visualisation and explanation to help an audience understand what was found and why it matters. A good report states the question, source of data, method, result, limitations and recommended action. It distinguishes facts from assumptions and avoids claiming more certainty than the evidence supports.

In Nigeria, data science can support crop-yield planning, disease surveillance, credit risk, fraud detection, traffic management, telecommunications, educational planning, energy demand, public budgeting and logistics. Effective solutions must consider local language, infrastructure, connectivity, representative data and the possibility that historical records contain social or regional biases.

Career paths include data analyst, business-intelligence analyst, data scientist, statistician, machine-learning engineer, data engineer, database specialist, research analyst and data-product manager. Useful foundations include statistics, Python or R, SQL, spreadsheets, visualisation, communication, ethics and knowledge of a chosen application domain. Students should build small, well-documented projects rather than collecting certificates without practical evidence.

Key Takeaways

Check Your Understanding

General Revision Questions

1. Explain the data science lifecycle and show why its stages are iterative.
2. Differentiate qualitative, quantitative, structured and unstructured data using suitable examples.
3. Discuss sampling bias and three dimensions of data quality.
4. Compare spreadsheets, relational databases and programming languages as data tools.
5. Explain how missing values, duplicates and outliers should be investigated.
6. Compare the mean, median and mode and state when each is appropriate.
7. Explain why correlation does not necessarily imply causation.
8. Differentiate supervised, unsupervised and reinforcement learning.
9. Discuss overfitting and the purpose of separating training data from test data.
10. Describe three Nigerian applications of data science and the ethical issues they may create.

Suggested Practical Exercises

• Create a spreadsheet containing 30 student scores, compute summary statistics and produce two suitable charts.
• Design a short questionnaire, identify its target population and explain how sampling bias could occur.
• Clean a deliberately untidy dataset containing duplicates, missing values and inconsistent categories.
• Use SQL or a database interface to create a simple table and retrieve selected records.
• Write a short Python program that calculates the mean, minimum and maximum of a list of values.
• Evaluate three published charts and identify any misleading design choices.
• Prepare a mini data report that includes a question, method, findings, limitations and recommendation.

Concise Glossary

Selected References and Further Reading

• Bruce, P., Bruce, A. and Gedeck, P. Practical Statistics for Data Scientists.
• Grus, J. Data Science from Scratch.
• McKinney, W. Python for Data Analysis.
• Provost, F. and Fawcett, T. Data Science for Business.
• VanderPlas, J. Python Data Science Handbook.
• Wickham, H. and Grolemund, G. R for Data Science.