GLAM/Newsletter/August 2026/Contents/USA report

This Month in GLAM – Volume XVI, Issue VIII, August 2026

USA report

GLAM USA Report (Past & Upcoming)

WikiBento: Roll our your own GLAM stats dashboard

Have you ever wanted to make your own dashboard to track your GLAM wiki or Wikimedian in Residence assets? A new project WikiBento allows you to do this and more. It is now open in alpha testing mode for those in the GLAM wiki community to modify and create their own dashboards.

The Problem: The observabilty of Wikimedia content and its usage has always been tough to track. We have a diverse range of metrics and reporting tools from Magnus Manske, the WMF, or other contributors, but they are not unified in any single view. And some faciliites like the Wikimedia Commons Impact Metrics API have never had an official user interface at all. This inadequate measurement and evaluation tools have been highlighted by the GLAM wiki community for many years.

The Solution: With WikiBento, users can display statistics, top 10 lists, SPARQL queries, individual Wikipedia articles or large image galleries, all in a customized view that can be saved and shared. Inspired by the rectangular compartments of Japanese bento boxes, WikiBento reads in a configuration of widgets and actions from a JSON file that can be generated from the authoring interface. After the user configures what they want, WikiBento can output the JSON file so users can store it on-wiki to be loaded again in the future.

Examples


Screenshot	Summary
	A WikiBento board can explore a particular Commons category (ie. Images from Metropolitan Museum of Art) and display the statistics and content related to it. Each bento box corresponds to either an API call or a tool call, displaying its output. WikiBento GLAM statistics for Images from Metropolitan Museum of Art
	A project like WikiPortraits can use WikiBento to track not just a category, but track popular Wikipedia pages and their status, such as with the article about skater Alysa Liu: WikiBento exploration for WikiPortraits and 2026 Winter Olympics coverage
	WikiBento boxes can go beyond just displaying content: they can be interactive and feed info from one to another, so that you can set up a sophisticated chain of events (currently a work in progress). Here is an example where you can select different GLAM institutions, and get the stats to change: Basic button interaction on WikiBento to switch GLAM institutions
	Ever had a bunch of Wikidata SPARQL queries you want to keep an eye on? You can configure each WikiBento box with a different one and show them in a dashboard. In fact, you can even query Wikimedia Commons using the third-party QLever endpoint and get statistics on Commons files: Multiple SPARQL queries in WikiBento

Future work

WikiBento is a work in progress and was initially built for a specific need – to provide a unified way to display output from the many tools in GLAM wiki work, while providing a much needed interface for API calls. As work on the dashboard function progressed, more features started to emerge. Connecting widgets provided a powerful way to chain capabilities together, providing a way to experiment with interactive multimedia.

To try it out, users can load one of the examples above and then click on the "gear" icon to change parameters – set it to a category you are interested in, or change the article to the one you care about. The idea is to experiment, drag, or reconfigure it for your uses.

WikiBento started as a simple dashboard but has grown to be something more. It is simultaneously:

A Grafana dashboard
An interactive Pinterest board
A clickable HyperCard-like interactive space
A Yahoo Pipes workflow.

So is it an activity monitor? A content platform? A multimedia browser? An exploratory storytelling experience?

You don't need to choose. It can be all of these, in any combination. That's what we are trying to find out and experiment with.

I also invite others to try it and provide feedback on the project at its Github repository, or contact Andrew Lih (User:Fuzheado) directly.

Addendum: This project was made possible with the coding assistance of an agentic coding system: Pi agent and Deepseek V4. More details can be found at the github repository.