Data Science for All : A Foundation Course

Identify data types and why they matter for storage and processing. Explore qualitative data, including nominal and ordinal subtypes, with examples like gender, color, country, and rating.

Distinguish discrete and continuous quantitative data, and explore ordinal versus discrete, with examples like price and temperature, plus other data types such as text, image, and graph data.

Explore data collection methods, including internal data from website interactions and offline visits, surveys and net promoter score, and external data from open source data such as public records.

Explore data storage solutions by location and data type, including distributed and cloud storage, and distinguish tabular data from unstructured data in relational and document databases with SQL and NoSQL.

Discover how data pipelines automatically extract, transform, and load data from multiple sources into storage, and monitor data quality to ensure clean, usable data.

Learn data preparation as the second phase of the data science workflow, transforming data into fixed columns, standardizing formats, applying missing value treatment, and assigning unique identifiers.

Act as a data scientist in a retail company to determine total purchases for a specific day using data rules and the provided data table.

Practice applying sentiment analysis to product reviews, and address missing product mappings when reviews lack corresponding product names to ensure accurate sentiment attribution.

Perform exploratory data analysis (eda) with descriptive statistics and univariate analysis to understand data, reveal relationships, and form hypotheses from Titanic survival patterns by gender and class.

Practice counting data points using descriptive statistics to determine how many values exist in a dataset, with 60 values in this example.

Practice analyzing the 100m world record data, noting how track surfaces, shoe design, and training methods drive improvements, and improve charts with axis labels and appropriate y-axis range.

Learn a five-step data science approach to hypothesis testing, from framing a question to testing a null hypothesis, significance, and interpreting results on conversion-rate experiments.

Practice a social media campaign analysis by testing the clickthrough rate as a hypothesis, outlining metrics, sample size, the test, and significance with ethics in mind.

Train a supervised learning model on input-output pairs to predict unseen data, with examples like spam filtering, speech-to-text, online advertising, and price prediction. Evaluate with train-test splits and accuracy metrics.

Identify supervised machine learning problems by evaluating examples like predicting cricket world cup winners and recommending books from past purchase history, while distinguishing hypothesis testing scenarios.

Choose the model with higher precision for coupon redemption; although both have 90% overall accuracy, algorithm one delivers higher precision (83.3%) than algorithm two, making it the right choice.

Explore unsupervised learning, where models uncover patterns in input data without labels, with examples like customer segmentation, fraud detection, image segmentation, and movie recommendations from the Movie Magic case study.

Practice identifies unsupervised machine learning by distinguishing grouping tasks from supervised predictions, using past data to forecast demand and classify items like cat or dog images.