Skip to content
Shyenova

GIS & Spatial Analysis

Decision Maps

For every view on the map, name the decision it triggers and the person who makes it. Views that fail the test get cut.

Most spatial dashboards are built to display data rather than to trigger action, and the difference determines whether anyone opens them twice. A map with fifteen toggleable layers and no default state asks the user to construct their own analysis; a map that opens on the six sites breaching threshold this week does the analysis for them.

The design discipline is narrow and unforgiving: every view must trace to a decision somebody makes and a threshold at which they make it. Applied honestly it usually cuts the requested scope by half, which makes the remainder considerably more likely to be used.

How the work runs

Four phases, and what comes out of each.

Durations are indicative and assume reasonable data availability. Where data is poor we say so in the first fortnight rather than modelling on sand.
01

1–2 weeks

Start from the decisions

We inventory the decisions the tool is meant to improve, who makes them, how often, on what information, and at what threshold they act. Requirements that do not trace to something on that list do not survive. This is the phase that determines whether the thing gets used, and it is the one most often skipped in favour of a layer list.

Outputs

  • Decision inventory with owners, frequency and action thresholds
  • Views specified against decisions, with the rest explicitly cut
  • Success measure agreed before anything is built
02

2–3 weeks

Prepare the data for the browser

Web performance is mostly determined before any interface code is written. Geometry is simplified per zoom level, cut to vector tiles, indexed and cached so that large datasets render quickly on ordinary connections. A dashboard that takes eleven seconds to load is a dashboard nobody waits for, however good the analysis behind it.

Outputs

  • Vector tile pipeline with zoom-appropriate generalisation
  • Spatial indexing and caching strategy
  • Performance budget set and measured against real connections
03

4–6 weeks

Build the views that matter

Built on open source (MapLibre, PostGIS, standard web tooling) so there is no per-seat licence and no ceiling on who can be given access. Views open on the exception rather than on everything, work on a phone for field users, and present a recommended action alongside the map rather than leaving interpretation to the reader.

Outputs

  • Exception-driven views deployed, mobile included
  • Alerting tuned to actionable thresholds
  • Role-based access with the field and office views differentiated
04

2–3 weeks

Measure use and cut what is not used

After a period of live running we look at what is actually opened and what is not, and remove the latter. Dashboards accrete; keeping them useful means deleting things. We hand over with the analytics in place so this can continue without us.

Outputs

  • Usage analysis against the success measure agreed in phase one
  • Views removed or reworked on the evidence
  • Handover, documentation and operator training

What good looks like

The ranges we design toward.

These are the improvement ranges these levers are known to produce where the underlying practice has not been examined recently. They are targets to test against your baseline, not promises.

30–50%

Scope cut

Typical reduction in requested views once each is required to trace to a named decision.

< 2s

Load target

Performance budget we design to on an ordinary mobile connection, measured rather than assumed.

0

Per-seat licences

Built on open source, so giving the whole field force access costs nothing extra.

Deliverables

What you are left holding.

All of it handed over, documented, and with a session for whoever inherits it.
  1. 01Decision inventory that determined what got built and what did not
  2. 02Deployed web GIS on open source, no per-seat licensing
  3. 03Mobile-capable field views with role-based access
  4. 04Vector tile pipeline meeting a stated performance budget on real connections
  5. 05Usage analytics and a documented process for cutting what nobody opens

What an engagement looks like

A representative shape.

Illustrative. This is the shape of work in gis & spatial analysis, not an account of a specific client. We do not publish named results, and we would be wary of a firm that publishes them without the client in the room.

10 weeks

  1. 01Situation

    A distribution business was assessing eight candidate locations using ten-kilometre radius catchments, and had recently opened two sites after which nearby existing sites declined without anyone having forecast it.

  2. 02What we did

    Geocoded three years of transaction data to customer origin, built drive-time catchments calibrated against where customers demonstrably came from, and fitted a gravity model to allocate demand between competing sites.

  3. 03Where it landed

    Modelled catchments were substantially smaller and differently shaped than the radii implied. Two candidates that had ranked highly on gross population fell sharply once cannibalisation was subtracted, and the ranking changed at the top.

Where this connects

Rarely a self-contained problem.

These are the adjacencies that come up most often on this kind of work. Naming them is usually cheaper than discovering them halfway through.

Questions

The ones we get asked.

Why open source rather than ArcGIS Online or similar?

Mostly because per-seat licensing changes behaviour in the wrong direction. When each additional user costs money, organisations ration access and the people who most need the map (field staff) are the ones who do not get it. Open source removes that constraint. The commercial platforms are good products and if you already have enterprise licensing the calculus differs, but we have no stake either way.

Will this work offline in the field?

Partially, and it is worth being precise about what that means. Cached map tiles and recent data work offline; live queries and anything requiring current data do not. For genuine offline-first field collection there are purpose-built tools we would point you at rather than rebuild. We usually design for intermittent connectivity, which is the realistic condition, rather than full offline.

How many layers can it handle?

Technically a great many, which is the wrong question. The constraint is cognitive, not computational, a map with twelve toggleable layers and no default state gets abandoned. We would rather build four views that each open on a specific exception than one viewer that can display everything and therefore recommends nothing.

Can field staff collect data through it as well as view it?

Yes, point capture, photo attachment, form submission and validation on entry all work on a phone. Where this matters most is closing the loop on monitoring: an alert generated from satellite detection gets verified in the field and the verification feeds straight back, which is what keeps the detection improving instead of decaying.

Start a conversation

Talk to us about decision maps.

Most useful first conversations are thirty minutes and involve no proposal. Describe the problem you are looking at and we will tell you honestly whether it is one we can help with.

Or write directly to hello@shyenova.com