How farishelmi17/applied-data-science-capstone-spacex works
Main Components
• Data Collection Modules – Represent the two distinct data collection approaches:
○ API-based Collection (from notebook:Data_Collection_yJPxhv2oU.ipynb)
○ Web Scraping (from notebook:Data_Collection_with_Web_Scraping_nI89VIRCE.ipynb)
• Data Wrangling – Represented by the notebook dedicated to cleaning/transforming data (notebook:Data_Wrangling_9HnvfsJ5G.ipynb)
• Exploratory Data Analysis (EDA) Modules – Split into at least two parts:
○ SQL-based Analysis (notebook:Exploratory_Data_Analysis_with_SQL__eqznon1EA.ipynb)
○ Visualization Lab (notebook:Exploratory_Data_Analysis_with_Visualisation_Lab_jJkKVG6F1.ipynb)
• Interactive Visual Analytics – Utilize Folium for geographic and interactive mapping (notebook:Interactive_Visual_Analytics_with_Folium_M8uUhCmHY.ipynb)
• Predictive Analysis – The machine learning pipeline for predicting landing outcomes (notebook:Predictive_Analysis_-_Machine_Learning_Lab_hdUi_lnX5.ipynb)
• Dashboard Application – A Python Dash app (spacex_dash_app.py) that likely integrates outputs from the predictive model and visual analyses.
• Final Presentation – The PDF slide deck (SpaceX_compressed.pdf) which summarizes findings and insights.
Relationships and Data Flow
• The cleaned data branches out into EDA modules (both SQL-based and visualization lab notebooks) and also feeds into the Predictive Analysis pipeline.
• After model training and prediction, results flow into the Dashboard Application, providing an interactive interface for users.
• Additionally, the final presentation is built on the insights derived from the entire pipeline.
Architectural Patterns and Design Principles
• Pipeline Architecture – Clearly delineate steps in the end-to-end machine learning workflow. Each phase (collection, wrangling, analysis, prediction, presentation) is a distinct module interconnected in sequence.
• Separation of Concerns – Notebooks focus on individual tasks (data collection, wrangling, etc.), while the Dash app handles visualization, ensuring clarity.
• Modularity – Treat every notebook as an independent component that feeds into a larger, cohesive system.
• Integration of External Technologies – Indicate use of Python and libraries such as Folium, SQL tools, and Dash. Note that even if explicit database setups aren’t visible, the EDA with SQL suggests a backend data storage or querying component.
• Every component has a clear label.