Data Science

What is Data Science?

Data Science uses programming, statistics and domain knowledge to extract useful insights from data. It can help explain past trends, support decisions and build predictive models when appropriate.

• Compiler: Python is a programming language widely used for data analysis, automation and machine learning.
• Header file: Provides declarations for functions and types.
• Function: A reusable block of code that performs a task.
• Use cases: Business analytics, forecasting, customer insights, scientific research and many other data-driven fields.

Basic workflow

Ask a question → collect data → clean and explore it → analyse or model → communicate findings.

Data Science fundamentals

A Data Science workflow often starts with a question and a dataset, then moves through cleaning, exploration, analysis, modelling and communication.

#include <stdio.h>

int main(void) {
    char name[40] = "Ravi";
    int age = 21;
    printf("Hello %s, age %d\n", name, age);
    return 0;
}

Practise loading a CSV, checking column types, summarising values and handling missing data in a Jupyter Notebook.

Python for Data Analysis

Reusable functions help keep analysis steps clear, repeatable and easier to test.

#include <stdio.h>

int add(int a, int b) {
    return a + b;
}

int main(void) {
    printf("Sum: %d\n", add(4, 6));
    return 0;
}

• Declaration: A DataFrame stores tabular data in rows and columns.
• Definition: A function groups steps into a reusable operation.
• Parameters: Parameters let a function work with different inputs.
• Recursion: A notebook records code, outputs and explanatory notes in one workflow.

Exploratory Data Analysis

Exploratory Data Analysis (EDA) uses summaries and charts to understand distributions, spot anomalies and explore relationships before drawing conclusions.

#include <stdio.h>

int main(void) {
    for (int i = 1; i <= 5; i++) {
        printf("%d\n", i);
    }
    return 0;
}

Use describe(), value counts, histograms and scatter plots to inspect data. Check whether missing values or outliers may affect conclusions.

Pandas, NumPy and Visualisation

NumPy supports efficient numerical arrays, while Pandas provides convenient tools for tabular data. Visualisation libraries help communicate patterns clearly.

#include <stdio.h>

int main(void) {
    int score[] = {70, 85, 92};
    int *ptr = score;
    printf("First score: %d\n", *ptr);
    return 0;
}

Practise array indexing, string handling, address-of (&) and dereference (*) operators. Avoid accessing memory outside array bounds.

Statistics essentials

Statistics help summarise data and quantify variation, patterns and uncertainty.

#include <stdio.h>

struct Student {
    int id;
    char name[30];
};

int main(void) {
    struct Student s = {101, "Asha"};
    printf("%d: %s\n", s.id, s.name);
    return 0;
}

Review central tendency, spread, probability, sampling and the difference between correlation and causation.

Machine learning basics

Machine learning uses data to fit models for tasks such as prediction or grouping. A good workflow includes validation and careful choice of evaluation metrics.

#include <stdio.h>
#include <stdlib.h>

int main(void) {
    int *values = malloc(3 * sizeof *values);
    if (values == NULL) return 1;
    values[0] = 10;
    printf("%d\n", values[0]);
    free(values);
    return 0;
}

Understand supervised learning, including regression and classification, and compare model results using appropriate metrics.

File handling

Data import and export allow analysts to work with CSV, Excel and other common formats. Validate schemas, missing values and data types after loading.

#include <stdio.h>

int main(void) {
    FILE *file = fopen("notes.txt", "w");
    if (file == NULL) return 1;
    fputs("Practise C every day.\n", file);
    fclose(file);
    return 0;
}

Practise reading CSV and Excel files with Pandas, exporting cleaned data and documenting the steps used to transform it.

Testing and project workflow

1. Define the analytical question and success criteria.
2. Validate source data and document assumptions.
3. Check missing values, duplicates and unexpected values.
4. Inspect intermediate tables and chart the most important patterns.
5. Record methods, limitations, metrics and recommendations.

Prefer reproducible notebooks, reliable validation, clear charts and honest explanations of uncertainty.

Data Science glossary