G

GeoWill

@GeoWill·17 public vizzes

Loading thumbnail…

copy to temp table method for updating addressbase

This example demonstrates a PostgreSQL method for efficiently replacing the contents of a large AddressBase database table by loading incoming CSV data into temporary tables, building indexes and constraints there, and then using a transactional rename swap to minimize downtime. The script first copies data from `addressbase_cleaned.csv` and `uprn-to-councils.csv` into empty temp tables, then adds primary keys, GiST spatial indexes on `location`, and B-tree indexes on `postcode` and `uprn`. Within a `BEGIN`/`COMMIT` block, it drops foreign keys, renames the existing tables to `_old`, promotes the temp tables to the main names, and re-adds the foreign key constraints. Timings for each operation illustrate the relative cost of loading, indexing, and swapping.

Loading thumbnail…

Overly complicated way of "pick a postcode that is a) inside the (multi-)polygon and b) far away from edges"

A static map highlights postcodes within a chosen administrative boundary by rendering the polygon's internal buffer and displaying point locations that fall inside it. The SQL query first simplifies the largest polygon part, then computes a buffer distance based on the average bounding box dimensions. It uses PostGIS functions like `st_buffer`, `st_simplify`, and `st_within` to filter UK postcode points from the `uk_geo_utils_onspd` dataset, returning a single matching postcode and its geometry. The result shows only the buffered region and the selected point, not the original boundary.

Loading thumbnail…

Gist 737374b53aa8f7261b66437904884375

This gist presents a Keybase proof of identity, demonstrating a cryptographic link between the GitHub account "geowill" and the Keybase user "willroper." The document serves as a public, auditable claim of ownership, containing a signed JSON object with the user's public key and a merkle root. The signature was generated using the keybase command line client and verified through the displayed signature block. The gist outlines the command-line process for verification and references the Keybase API for identity lookups.

Loading thumbnail…

Append some new json to some existing json

This example shows how to merge two JSON files into a single, updated JSON object. The command-line tool `jq` uses its `-s` flag to read both input files into a single array, then applies the `.[0] + .[1]` filter to concatenate the two objects, writing the result to `updated.json`. The process relies entirely on `jq`’s stream editor and its built-in addition operator for objects. The source data originates from the two input files, and the output is a new file.

Loading thumbnail…

QGIS Function to make strings ascii only

This example demonstrates a custom QGIS function that converts place names into lowercase ASCII-only strings, replacing spaces with hyphens and translating accented characters like ß, à, and ñ to their closest ASCII equivalents. The function, defined with the `@qgsfunction` decorator and grouped under 'Custom', uses Python’s string methods `lower()`, `replace()`, and `translate()` with a manually defined `ascii_dict` mapping Unicode code points to replacement strings. It is designed for use in QGIS expression engines, taking a string input and returning the cleaned version. The code also includes a commented-out alternative using `encode('ascii', 'asciify')`, but the active path relies entirely on the translation dictionary.

Loading thumbnail…

Simple shell script to merge shapefiles from INE

This example shows a shell script that downloads, reprojects, and merges Mexican electoral geography shapefiles from the INE’s distritación 2017 federal collection into a single output file. It processes layers such as `SECCION`, `ENTIDAD`, and `DISTRITO` for all 32 states, using `ogr2ogr` from GDAL to reproject each source into EPSG:4326 and append the features. The script handles a structural anomaly for file number 06 by adjusting the source path, and it relies on `/vsizip//vsicurl/` for remote zip access. The merged result is saved as a new shapefile named with the layer and `-wgs84` suffix.

Loading thumbnail…

Scripts to merge local electoral districts in Mexico

This example shows a shell script that merges 32 individual shapefiles of Mexico’s local electoral districts into a single national layer. The script uses GDAL’s `ogr2ogr` command-line utility, with two variants: one that reprojects each source file to EPSG:4326 (WGS84) during the merge, and another that preserves the original projection. It reads the shapefiles directly from zipped remote archives via the `/vsizip//vsicurl/` virtual file system, downloading each state’s `DISTRITO_LOCAL.shp` from the IFE cartography server. The first execution creates the output file, while subsequent iterations append data to it, building the combined dataset.

Loading thumbnail…

Gist c2315899ad3a0cad09304146add83d04

This visualization compares village codes from a JSON dataset against a Shapefile, showing which villages appear in one but not the other. It displays the mismatched codes in a text list, summarizing the total unique counts for each source (7,851 in JSON, 8,017 in Shapefile). The example uses a straightforward text rendering approach rather than graphical mapping, relying on file parsing and set operations to identify discrepancies. The data comes from GeoWill’s gist, with the output formatted as plain text lists for easy inspection.

Loading thumbnail…

Gist 1fded528ddc844c4a22e72d8d418ad70

This visualization compares village locations from a JSON dataset with those from a Shapefile, highlighting discrepancies. It plots the geographic points, likely using a mapping library to render the Shapefile boundaries and a separate layer for the JSON data points. The output is a text file listing 166 village codes present in the Shapefile but missing from the JSON, quantifying the mismatch between the two data sources.

Loading thumbnail…

Gist 88d08870d21ac3591ac52dfcb82a8329

This visualization maps towns and counties in Taiwan that are absent from the ethnical electoral regions dataset, using a force-directed layout to position labeled points. The code employs `d3.forceSimulation` for spatial distribution and `d3.geoPath` to render geographic boundaries, with data drawn from a CSV list of county-town pairs. It highlights gaps in regional coverage, focusing on municipalities in Yunlin, Lienchiang, Kinmen, Penghu, and Chiayi. Text labels are placed via SVG elements, and the rendering likely relies on D3’s projection utilities to align the points with the map.

Loading thumbnail…

villages

This visualization maps administrative divisions across Taiwan, showing village boundaries and their relationships within counties and townships. The code reads a CSV dataset of village codes and names, then uses d3.forceSimulation to position nodes representing each village, with links connecting villages sharing the same town or county. The rendering relies on SVG paths generated through d3.geoPath, while the underlying geographic data is sourced from a gist containing Taiwanese administrative boundaries.

Loading thumbnail…

villages

This visualization shows the distribution of villages across Taiwan, mapping each administrative unit to its geographic position. It uses D3's `d3.forceSimulation` to lay out points based on hierarchical township and county codes, with `d3.geoPath` rendering the underlying map boundaries. Village names are displayed in both Chinese and English, sourced from a CSV file containing official township and village codes. The visualization employs a color scale to distinguish counties, with labels positioned via force-directed collision avoidance.

Loading thumbnail…

towns

The visualization displays the geographic distribution of towns across Taiwan, using their names, county affiliations, and administrative codes from a CSV dataset. It presents a map-like layout where each town is depicted as a point or region in relation to its county, with labels showing traditional Chinese names and English romanizations. The code leverages d3.geoPath to render town boundaries or locations from the geospatial data, while d3.forceSimulation could arrange overlapping points into a readable pattern. The rendering relies on Canvas and browser APIs to draw the map, with the source data sourced from a gist by GeoWill.

Loading thumbnail…

Missing districts

The visualization shows which townships and districts in Taiwan are absent from a given dataset, with each missing area highlighted on a map. It uses a force-directed layout via `d3.forceSimulation` to position label points for each missing town, while county boundaries are drawn with `d3.geoPath`. The map is rendered on a Canvas, where `getImageData` detects whether a pixel falls within a county, enabling precise placement of labels. Data comes from a CSV listing county and town names, with the result revealing clusters of omissions, particularly in Yilan, Tainan, and Taoyuan.

Loading thumbnail…

Gist dd10fa64484c334e066db46d1b74959b

This example shows a geographic grid of vertical and horizontal line segments covering a small area near Shwatka, Yukon, generated from a what3words API request. The visualization uses a GeoJSON MultiLineString containing coordinates for 14 horizontal and 14 vertical lines, which together form a dense mesh of cell boundaries. The code likely renders these lines with a mapping library such as D3.js, using `d3.geoPath` to project and draw the geographic coordinates onto a canvas or SVG. The API response metadata confirms successful retrieval from what3words, with the grid representing a localized portion of their three-meter word-based location system.

Loading thumbnail…

San Juan Island Geology

This example displays a Mapzen basemap centered on San Juan Island, Washington, at zoom level 12. The visualization uses the Mapzen.js library, instantiating a map with `L.Mapzen.map` and specifying a scene file. A Mapzen geocoding search box, created with `L.Mapzen.geocoder`, is added to the interface, and the zoom control is repositioned to the top right using `map.zoomControl.setPosition`. The map itself is rendered in an HTML page that loads Mapzen’s CSS and JavaScript resources.

Loading thumbnail…

One Minute Map

This example demonstrates a one-minute map centered on San Francisco, showing an interactive basemap with search and location-finding controls. Built with the Mapzen JavaScript library and Leaflet, the visualization uses the Mapzen Refill basemap style to render the geographic data. A geocoder search box in the top-right corner allows users to find places, while a locator button in the bottom-right identifies the user’s current position. A hash feature enables deep linking to the map’s coordinates and zoom level, making the map state shareable via URL.