Skip to content

What product is Security Force Monitor making? (July 2017)

Tom Longley edited this page Jul 27, 2017 · 1 revision

The Product is …

This product is a publicly-accessible website that provides data about commanders and other personnel, command chains and other linkages, locations and geographical areas of responsibility of police, military and other state security forces from different countries all tracked through time using public and open sources. It is also a research tool for Security Force Monitor, and the first of a set of products and tools that Security Force Monitor may create in pursuit of its mission.

The Product is built to serve three main user groups…

First, people in US and others governments who are tasked with monitoring sanctions regimes and compliance regimens regarding security assistance (in particular the Leahy laws).

Second, country-experts and security force researchers in international and large national human rights organizations.

Third, the SFM research team, by providing an effective platform for the capture, storage and analysis of data.

The behaviour changes that The Product drives is …

A challenge to assumptions in guests that detailed information about identities of perpetrators and units is not available. Countering a trend in human rights reporting to not identify specific perpetrators for abuses committed by security forces.

The Product will be a success when ….

Both sets of guests expect The Product to know more than they do; routinely check the site when investigating people and organisations alleged to be involved in human rights violations; de facto replace parts of their own research in these areas with work done by The Product; and, start citing The Product in their published works.

Who uses The Product?

The Product will meet the needs of three types of users:

  • A guest is anyone with a web browser who visits the site. Two groups of guests will most benefit from The Product, as stated above.
  • A staff researcher is a member of Security Force Monitor tasked with the day-to-day updating and improvement of the data. They could be permanent (staff), short term (interns), or potentially personnel from partner organizations.
  • An administrator is a member of the Security Force Monitor who is a task manager for staff researchers. They control who can do what on The Product, resolve problems with the platform’s use, and manage the overall data publication process

What do people need to do on The Product?

A guest on The Product

...needs to be able to:

  • Clearly understand what The Product does and whether it can be useful to them
  • Be assured of the privacy of their use of The Product
  • Use the site in their own language
  • Be able to establish quickly whether information about security forces in a specific country is in The Product
  • Forage around data about a particular branch of the security forces in a country
  • Learn a little general information about the different security forces that exist in a country
  • Find out whether a specific organization/unit or person is contained within The Product
  • Find out whether The Product contains data about persons, organizations and incidents that happens in a place (or area) at a specific time (or timeframe).
  • Clearly see the connection between any organization and the chain of command upwards and downwards
  • Learn who the command personnel for an organization are, and how this has changed over time
  • Find help about how to use the site quickly
  • See the career history of any Person contained in The Product
  • See the connections between any organizations, especially via memberships
  • Advise us if they have problems using the site or if they know something we don’t
  • Browse all the source information that was used to create any datapoint contained in The Product
  • Figure out why they should rely on the data contained in the The Product, by learning about its quality/confidence score
  • Download data from The Product for a specific record, groups of records, or every record, in a way they can use
  • Get a well laid-out print-out of any record on The Product
  • See potential connections between units and/or people and incidents
  • Understand when data was last updated
  • Find the sources for our data
  • Understand SFM’s confidence in its own data

… would think it was cool if they could:

  • See where data on persons and organizations overlaps with data about incidents, and get a copy of all the data and source used to make this analysis
  • See connections between incidents and people and organizations - including “up” the command chain
  • Make lists of people, organizations and places that are already in the database
  • Learn when there are updates to those records
  • Explore selections of records on a map
  • Contribute data to The Product, or review existing data-points
  • Upload their own incident data and augment it with the data in The Product
  • Use visuals from The Product to quickly incorporate into their own report

A staff researcher on The Product

In addition to those things a guest needs to do, a staff researcher needs to be able to:

  • See all the data in The Product, not only that which is published (visible to guests)
  • Add, update and remove persons, organizations, incidents and sources from the website
  • Add, update and remove connections between different organizations and persons
  • Easily add geographic information to records
  • Search, edit and remove sources from The Product
  • Review data added to the website by other staff researchers, prior to its publication
  • Easily find older, stale data as part of work to review and update the data in The Product

… would think it was cool if they could:

  • Create a shared private archive of any source to be automatically updated when new sources are entered
  • See contributions to the dataset made by themselves and other staff researchers
  • Being able to track down modifications and updates
  • Flag up any issues in the data in The Product
  • Assign themselves or other staff researchers to fix them
  • See a list of current issues in the dataset, and who is assigned to fixing them
  • Add, update or remove a lot of data at once
  • Pick up items from a research queue created by a crawler, and review machine-populated statements about the data in those sources
  • Create an SFM-wide research queue for all sources, digital and nondigital
  • Track status of sources within the research queue

An administrator of The Product

In addition to things a guest and staff researcher needs to do, an administrator of The Product needs to be able to:

  • Examine how the website is being used and by whom
  • Troubleshoot problems experienced by anyone using the site
  • Add, update, suspend and remove user accounts
  • Control the public visibility of a specific data point or selections of data, publishing or un-publishing as required
  • Add, update and remove non-data pages from the website (for example FAQs, or help pages)
  • Take the website offline in the event of problems
  • Review work of staff researcher and provide comments for review

Qualities of The Product

The Product should have the following general qualities:

  • Integrated experience for all users: there is little need to separate out the ‘back end’ and ‘front end’ of The Product. Viewing data, adding data and other research and administrative tasks will go on through the same application.
  • A simple, modern civic tech service: The Product should look and feel at home with sites like WhatDoTheyKnow and the CouncilMatic family of sites. It should use the basic conventions of the web, and not upend the expectations and web-use habits of guests. It should be easy for anyone to see what they will get by using the site.
  • Browse, search and filter: The design of The Product must be aimed at answering guests’ questions in the fastest and most effective way, with as little manual labour as possible.
  • Localized throughout: the interface, and where possible the content, of The Product should be rendered in the guest’s choice of language.
  • Responsive design: The Product should be straightforward for guests to use on handheld devices of their choice.
  • Minimal, appropriate, effective visualisation: The Product should include simple, familiar data visualization that provides key insight in the most distinctive and usual (and hence confusing) aspects of the dataset.
  • High quality data, high assurance service: Net platforms come and go, have outages and so on. Guests of The Product should always be able to to take data home.
  • Indexable, archivable content: The Product is a public resource built with philanthropic resource. Its content should be indexed by search engines and online archiving systems.
  • Takes advantage of machine cognition: The Product will make smart use of machine-assisted research techniques, progressively incorporating them into the process of expand and maintain its content, and in providing analysis for its users.
  • “Friendly” to low bandwidth: Guests with limited connectivity should be able to use the site to answer basic research questions

Development roadmap

The development of The Product starts from the initial cut on the “CMS” and May 2017’s round of user research. Much of the functionality is already present in the product so far, and many of our initial needs are to do with improving the layout.

Chunk 1: reboot and re-release

The vision for this chunk of work is to consolidate the strands of work (“CMS”, “Frontend”, “API”) into a single platform which provides what guests and the SFM team need. It’s focussed on putting out a far simpler version of the tool for guests to use, for us to crunch data into:

  • Alterations to site structure, navigation and visual look/feel
  • Updates to search and filter views, introducing more facets and a time-bounded search (specific date ranges, and timeframe - “last month”)
  • Redesigned record views presenting key record information, and incorporating simple map and ‘line of sight’ chart, and simple “all sources” section (rather than per data value)
  • Incorporation of basic in-line help functions for guests
  • Alteration of data model to render sources a distinct entity in the database
  • Improved and expanded data entry wizard, including absurdity checks, methods for adding and checking geographies, and fuzzy (probabilistic?) data validation checks to reduce possibility of entering duplicate persons and organizations
  • Data export functions for guest users
  • Introduction of user permission model and basic capability grouping for guests, staff researchers and administrators
  • For administrators, basic visibility controls for datapoints: publish, unpublish
  • Incorporation of non-data pages into the site, such as explanatory panels, FAQs, Support.

Chunk 2: feature extension

This vision for this chunk of work is to refine the user experience, introducing betters ways for guests to examine the integrity of information (sources, confidences), arranging data ad-hoc ways and pushing info towards guests:

  • UX improvements regarding how users view and access data quality and integrity information (source and confidence toggles)
  • Improved timestamped user action logging functions (create, update records etc) that is searchable by staff researchers and administrator
  • Development of “collection” feature to enable guests to make ad-hoc groupings of data and be alerted when those data-points have been updated (a notification system)
  • Improvements to display of data, including a map view for collections of data
  • Development of review functionality for data entry, including a “four eyes” system for all data
  • Update to data entry process to make an archive of sources, which is hosted on The Product, but will initially not be available to guests.
  • Archive of all sources, integrated into web product, searchable by both metadata and links to data-points

Chunk 3: machine-enhancement

This chunk is mostly to be decided and as at July 2017, Security Force Monitor is exploring the use of NLP and entity relation techniques to assist with data creation. This vision for this chunk is the incorporation of of machine-enhanced research workflows, and relevant features for inclusion in The Product.