Data Analysis, Geospatial Mapping & Interactive Dashboard
This project analyzes the condition and spatial distribution of Public Street Lighting (PJU) across Tasikmalaya Regency. The workflow combines data cleaning, exploratory analysis, condition classification, spatial joining with district boundaries, static and interactive mapping, and a Streamlit dashboard for easier exploration.
The raw dataset contains 5,589 PJU records with location, coordinates, condition, PJU type, pole type, and lamp type. After handling missing values, 4,235 records remained for condition analysis. A spatial join with the Tasikmalaya Regency district GeoJSON retained 3,229 records whose coordinates fall inside the mapped district boundaries.
Data Loading → Data Cleaning → District Extraction → Exploratory Feature Importance → Condition Classification → Coordinate Parsing → GeoDataFrame Creation → Spatial Join → District Summary → Static & Interactive Maps → Streamlit Dashboard
After non-spatial cleaning, the dataset contains 3,777 PJU in good condition and 458 damaged PJU.
Sukahening recorded the highest damaged count in the district summary shown in the notebook, with 31 damaged PJU and a damaged percentage of about 68.9%.
Among the technical attributes examined, lamp type showed a stronger relationship with PJU condition than pole type in the exploratory feature-importance analysis.
The notebook concludes that maintenance attention should prioritize vulnerable pole and lamp categories, while the findings should be treated as an initial indication because the available variables are limited and the category distribution is uneven.
The analysis is a snapshot based on the available attributes. It does not include factors such as installation age, installation quality, weather, accident exposure, or usage intensity. Because some categories have much smaller populations than others, percentage-based damage comparisons should be interpreted carefully.