Kepler.gl Beginner Help

[ChiefSosa1234]

We used Kepler.gl for my architecture class to visualize data for our building sites such as traffic data, activity centralities, surrounding building categories (residential, education, food, healthcare, etc.), heatmaps of pedestrian activity, etc. My professor never really told us how we can extract this information ourselves. We basically made these kepler maps from CSV files he fed us, we never learned how to do it ourselves. Is anyone able to help me out with understanding how I can teach myself to extract this type of data from the internet? Not sure where to even start.

[igorDykhta]

Hi @ChiefSosa1234

I suggest learning how to work with data in DuckDB. If you have a dataset, or several datasets that aren’t directly connected, you can use free AI tools to guide you through the basics and how to join datasets together, depending on what you want to do with your data.

If you have something more concrete in mind, ask here and we’ll try to figure out how to prepare your data.

DuckDB quick recipe: GeoJSON → filtered CSV with centroids

• 1) Setup (once)

INSTALL spatial; LOAD spatial;
-- Optional if reading via https:
INSTALL httpfs; LOAD httpfs;

• 2) Read your GeoJSON

-- Local file or an https URL
CREATE OR REPLACE TABLE g AS
SELECT * FROM ST_Read('data.geojson');

• 3) Keep only the part you care about

-- Example: keep a few categories
CREATE OR REPLACE TABLE g_sub AS
SELECT *
FROM g
WHERE amenity IN ('school','restaurant','hospital');

-- (Optional) clip to an area (WGS84 lon/lat)
-- CREATE OR REPLACE TABLE g_sub AS
-- SELECT *
-- FROM g
-- WHERE ST_Intersects(
--   geom,
--   ST_GeomFromText('POLYGON((-73.99 40.73, -73.98 40.73, -73.98 40.74, -73.99 40.74, -73.99 40.73))', 4326)
-- );

• 4) Compute centroids and export to CSV for kepler.gl

COPY (
  SELECT
    ST_X(ST_Centroid(geom)) AS lon,
    ST_Y(ST_Centroid(geom)) AS lat,
    name, amenity
  FROM g_sub
) TO 'pois.csv' (HEADER, DELIMITER ',');

Notes

• For polygons, ST_Centroid is fine; if you need a point guaranteed inside the polygon, use ST_PointOnSurface(geom).
• Add more filters (e.g., WHERE name IS NOT NULL or LIMIT 5000) to trim the output further.