https://github.com/do-me/semantic-hexbins

A light-weight demo app for geospatial semantic search. Designed for text data with geospatial references.
https://github.com/do-me/semantic-hexbins

hexbins leaflet semanticsearch transformers

Last synced: 22 days ago
JSON representation

A light-weight demo app for geospatial semantic search. Designed for text data with geospatial references.

• Host: GitHub
• URL: https://github.com/do-me/semantic-hexbins
• Owner: do-me
• License: mit
• Created: 2023-11-18T16:10:38.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2025-04-13T08:15:54.000Z (22 days ago)
• Last Synced: 2025-04-13T13:15:37.401Z (22 days ago)
• Topics: hexbins, leaflet, semanticsearch, transformers
• Language: JavaScript
• Homepage: https://do-me.github.io/semantic-hexbins/
• Size: 193 MB
• Stars: 2
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Semantic Hexbins

A light-weight demo app for geospatial semantic search. Designed for any kind of textual data with geospatial references.

Building on previous research:

- [An Application-Oriented Implementation of Hexagonal On-the-fly Binning Metrics for City-Scale Georeferenced Social Media Data](https://isprs-archives.copernicus.org/articles/XLVIII-4-W7-2023/253/2023/)

- [Developing a Privacy-Aware Map-Based Cross-Platform Social Media Dashboard for Municipal Decision-Making](https://isprs-archives.copernicus.org/articles/XLVIII-4-W1-2022/545/2022/)

Repository for the paper: XXX (still to submit)

![](screenshot_overview.png)

## Idea 

The paper describes an approach to use semantic similarity for geospatial purposes, like georeferenced social media data.

## Data samples

Ranging from 8 - 32 Mb for individual posts or 0.8 - 5.1 Mb for aggreagted posts, see data folder.

## Scripts

Scripts for data processing can be found here:

- https://gist.github.com/do-me/d60ea47d0dc97ba40c9d727bf26f7a77

- https://gist.github.com/do-me/dc8877049c2c074df3c7d8e707adf138

- https://github.com/do-me/fast-instagram-scraper

## Example Queries 

See the screenshots folder for query comparisons between the location-averaged and individual embedding indice.

## Performance 

Tested devices: 

- Windows laptop with Intel i7-8550 CPU

- Ubuntu laptop with AMD Ryzen 7 PRO 6850U

- Android phone Samsung S9 with Exynos 9810

- Apple iPhone 15 Pro with A17 Pro

Run times for a full layer update are significantly below 200ms with ~60ms inferencing time. Iphone 15 Pro averages 51ms (24ms for inferencing), averaged for 100 runs.

For comparison to a simple full-text search in JS see the standalone html file `benchmark_full_text_search_no_index.html`. It benchmarks dummy data in social media style with 4 columns: lat, lon, location ID and text.

![image](https://github.com/user-attachments/assets/9364866c-a1de-453c-a4b4-18f93fa6c549)

Screenshot results run on an M3 Max.