Caloom – overview

connecting the dots

reveal hidden patterns

data journalism, investigative reporting, news, education, research, forensics


Caloom 2009

Caloom started as a reaction to my research in energy and financial markets starting back in 2006. As research alone was not enough I turned my findings into something concrete. Due to what we saw was a possible economic slowdown I envisioned we should have a system where people could easily find useful resources nearby: version 1 of Caloom was born

2010 – 2013

In the years following map based services became all the hype. Caloom had already a technology that could create within minutes (I still have that code) most of the then available geo-services. So it had a solution to the problem many developers solved again and again. Selling this was not possible even potential customers would have loved to use Caloom.

2014 – 2015

Caloom Yle

Caloom AI

In the end Caloom evolved into a very smart, elegant and useful technology that would allow e.g. an online news site to interconnect it’s own content but also connect with “competitors” content while creating increased user value. Below are some posts and some older scribbles on Caloom but 99% is offline.

Caloom works as it searches through text for any what, where, when, who, and a value (number for example). This is structured and indexed. When texts are done in this way we can begin to cross reference them and search. We can then also look for connections that are very hard to discover by only reading the texts.

What is Caloom?

Caloom is a service that brings clarity in the intricate web of relations between connected and disconnected online textual content. This is done by  analyzing text, collecting the What + Where + When + Who + Worth (WX) from the text, structure the results with the WX data model and making this available.

Caloom offers easy to understand building blocks from which the user can explore the various narratives.Caloom shows what is the relationship between the different sources. Caloom paves the way for understanding the root causes and assist the human finding the answer on the final question: why

What does Caloom solve?

  1. Caloom connects non-connected data repositories. Imagine how one newspaper’s articles becomes connected to another newspapers’ article. How a person in one article appears in one or more other articles, from the same news paper or the other ones. What about podcasts, research databases and more sources getting connected. This without judgement, without trying to understand sentiments and context. Raw and pure links.
  2. Caloom unravels the often linear narratives and restructures them with the WX model. Once we stepped outside this one article (for example) and build bridges with other newspapers and textual sources, we can begin to show alternative narratives. The human memory is fallible, it creates bias and is selective for example. Now that we can discover and experience how a person has a long history of “doing A” and being in relationship with “B” we could review the current article in a different light perhaps.
  3. Caloom makes visible the relationships with other online content. Text can be beautiful and powerful but also tedious to read and understand. We combine text with visual aids to help you to discover narratives. This is only the beginning…

What opportunities does Caloom bring

  1. connect the content of the creators with one another
  2. Caloom brings clarity on how the content relates to other content, within the same domain or across domains
  3. visually discovering direct and indirect connections between textual content

How does Caloom do this all

  1. Caloom has a variety of front-end and back-end tools to assist content providers and their audience in finding, discovering, connecting and managing the WX based relationships
  2. Caloom uses a new linguistic algorithm we call WX, this algorithm uses What + Where + Who + When + Worth for extracting and analysing the text based content
  3. Caloom creates and stores the WX meta data and this is used to find relationships previously unseen across large sets of content

Who is our audience

  1. content providers
  2. educators
  3. decision makers

What technologies do we use?

Using the Frugal Innovation thinking we apply commonly available technologies and utilise them in the right context to come up with a solution that provides desired outcomes. Our focus is on MVPs and getting solutions out there, quick and working.

Python is our choice as it is in most cases the first API language supported for tools relevant to Caloom. Python is also an established language for data and text analysis. Caloom uses ElasticSearch, Celery, PostGIS and Django on the server side. Angularjs, jquery, d3 and html5 on the client side.

What do we have?

We have experimented a lot with an a base ecosystem that was meant to be a collaboration platform, but that experiment never materialised in a sellable product. This is visible (not so beautiful) and working at

Core of this ecosystem was the WX model and many re-usable tools.

Caloom tasks

  1. backend based on existing work
  2. front end based on bootstrap html5 template
  3. WX.js, a la ga.js
  4. WP plugin, WP has the largest marketshare among bloggers and content creators, 67%. Unlocking the independent media adds a powerful aspect.
  5. ODF plugin (see More and more governments are using ODF as their document format or have started to become compatible with ODF. It is an opportunity to get involved in e.g EU’s agencies which already support ODF

We at Caloom take an agnostic view on information, like a chemist looks at chemicals and builds models of how molecules and atoms are connected without a preference for one atom over the other. We scan the presented text and extract the words that fall into the WX model.

Use cases here