Lizmap dashboard: numbers right next to your maps

Why a Lizmap dashboard

Elected officials, managers and partners rarely read a map in detail. What they ask for often fits in a few numbers: cases per district, length of network renewed, area by land use, change from one year to the next. If those numbers are calculated separately in a spreadsheet, they quickly drift away from the map data, and you end up presenting two versions of the same reality in the same meeting.

A Lizmap dashboard solves this: indicators are calculated from the same layers as the map, on the same portal, with the same access rights. When the data changes, the number changes. The GIS specialist no longer has to rebuild a table for every request, and everyone looks at the same source.

• Tracking a program: progress of works, budget committed, features processed over the year.
• Reporting: key figures for a council meeting, a steering committee or an annual report.
• Spotting: districts where reports concentrate, equipment that keeps breaking down.
• Comparing: one year with the previous one, one area with another.

There are three tools, and they do not meet the same need: the Dataviz charts built into Lizmap, the dashboards module offered by LizPack, and MapBuilder, which lets users compose their own maps. The following sections cover them one by one.

Level 1: Lizmap's Dataviz charts

Lizmap Web Client can display charts without installing anything else. You configure them in the Dataviz panel of the Lizmap plugin for QGIS Desktop. According to the official Lizmap documentation, these charts are built on Plotly, the same engine as the DataPlotly plugin for QGIS, and nine types are available:

• scatter, bar, pie and histogram, for everyday needs;
• box plot, to compare distributions;
• two-dimensional histogram and polar chart, for finer analysis;
• sunburst, for hierarchical data;
• an HTML type, which shows calculated values in a template you write yourself, handy for a key figure.

Charts appear in a panel of the portal, placed according to the project setting: left, bottom or right of the map. They can also appear in information popups: for a layer linked to another by a relation, Lizmap can show, in the popup of a parent feature, a chart limited to its child features, for example the interventions on a single piece of equipment.

• Strengths: nothing to install, configured in QGIS, interactive charts, option to show a chart only when its layer is visible.
• Limits: each chart describes one layer; building a page of key figures, pivot tables and charts linked to each other needs something else.
• Ideal for: illustrating a layer with one or two charts, or enriching information popups.

The HTML type is worth knowing: it calculates values on the layer, for example a total or an average, and inserts them into a small template you write. You get a readable key figure with a sentence around it, and no chart at all. Combined with a few bar charts, it is often enough for a first, simple dashboard that answers the questions people actually ask.

Configuring a Dataviz chart step by step

The layer must first be published as WFS, with its primary key published in the layer properties: the Lizmap documentation lists this as a prerequisite. Then:

1. In QGIS, open the Lizmap plugin from the Web menu, then the Dataviz panel.
2. Add a chart and choose its type; the form adapts to the chosen type.
3. Give it a clear title and, if needed, a description: it accepts HTML for an explanatory sentence or a source.
4. Choose the layer, the X field and one or more traces of values.
5. For bar or pie charts, choose the aggregation: sum, average, count, median, standard deviation, minimum, maximum, first or last value.
6. Set whether the legend is shown and, if you wish, whether the chart appears in the popups of the parent layer.
7. Save the Lizmap configuration, save the project, then publish it again.

Advanced formatting (colors, axes, margins) goes through a layout setting in JSON format that follows the Plotly reference. Start without it: the default settings are readable, and you can refine the look later once the chart has proved useful to its readers.

Preparing data that makes good indicators

A dashboard is only as good as its data. The same flaws come back in almost every project, and they are fixed in the database, not in the chart:

• Consistent categories. Roads, roads and ROADS written three ways make three different bars. Use closed value lists in data entry forms, and clean up what already exists.
• Real dates. A date stored as text cannot be sorted or grouped by month. Use a proper date field.
• Explicit units. State in the title or description whether a figure is in meters, hectares or number of features, so that nobody has to guess what a bare number means.
• Views rather than copies. If an indicator needs a calculation (a join, a grouping), create a view in PostGIS rather than a copied table: the view stays up to date.
• Indexes. On a large table, an index on the filtered or grouped fields makes display much faster.
• Missing values handled. An empty field is neither zero nor a category: decide how to count it, and say so in the indicator's description.

If your data is still in files, now is the time to move it to PostGIS, where views, indexes and shared editing all become possible. The guide Publish a Lizmap map from QGIS explains how to publish PostGIS layers and why they avoid publishing the project again after every update.

Level 2: the dashboards module

For a real indicator page, LizPack, an open source geospatial infrastructure that hosts and maintains Lizmap, offers a dashboards module that adds to a Lizmap portal. Dashboards are composed in QGIS, then published next to their project. They combine:

• indicators, that is, key figures;
• charts;
• pivot tables;
• lists.

The elements filter one another: a selection in one narrows the others. They are recalculated on every display, from up-to-date data. So you prepare the dashboard where you already prepare the map, in QGIS, and publish it the same way, for example with the LizPack Publisher plugin.

Modules install on a running LizPack Map instance, on Lizmap 3.9. Each module states the Lizmap versions it works with.

The difference from Dataviz charts is one of scale: Dataviz adds charts to a map, while the module composes a whole page devoted to numbers, where the elements respond to each other. You can use both on the same portal: charts in information popups for detail, and a dashboard for the overall picture that managers and elected officials look at first.

Level 3: MapBuilder to compose maps on demand

MapBuilder meets a different need: letting your users compose a map themselves from the published layers, without QGIS, and save it. It is an open source module developed by 3Liz, the publisher of Lizmap, released under the MPL 2.0 license on GitHub.

• Combining projects. The module's documentation gives the example of an OpenStreetMap project and a local land use plan project brought together in one map.
• Printing. The composed map can be exported to PDF straight from the browser.
• Administration. The module is configured from the Lizmap administration.
• Following versions. Each MapBuilder version matches a Lizmap Web Client version: they are updated together, and the module README lists which version goes with which.

MapBuilder is useful when each department wants its own view of the same reference data without asking the GIS team for a new map. Combined with the dashboards module, it gives a portal where people check up-to-date numbers and build their own maps, while staying on the reference data.

MapBuilder does not create data: it reuses the layers already published on the portal. The quality of the maps people compose therefore depends on the quality of the original QGIS projects, their styles and their layer names. Well-named, well-organized reference data makes MapBuilder immediately useful; a project with cryptic layer names makes it confusing for everyone who opens it.

Setting up a Lizmap dashboard with LizPack Map

1. In the customer area, open the page of the running instance, Modules tab.
2. Choose the module: each one shows what it adds, its version and its documentation.
3. Click Activate: the module installs in under a minute, without interrupting the portal.
4. Compose the dashboard in QGIS and publish it next to its project.
5. Open the portal: the dashboard is available there, next to its project.

Deactivating removes the module from the instance without touching your projects or data. Added modules are part of the instance configuration, backed up and restored with it, so a restore brings back the dashboards along with the maps.

No instance yet? LizPack Map puts one online in about four minutes, with QGIS Server and a PostGIS database already connected, and a 30-day free trial.

Before announcing the dashboard, open it with an account from each user group concerned: check that the numbers display, that the filters respond, and that nobody sees data that is not meant for them. Ten minutes of testing here saves an awkward correction after the dashboard has been presented to a council or a steering committee.

The rules of a useful dashboard

A dashboard is not a catalog of every number available in the database or the spreadsheets of the department. The most useful ones follow a few simple rules:

• Start from a question. Where does the renovation program stand? makes a better dashboard than let us show the road data.
• Few indicators. Five to eight well-chosen elements get read; twenty get skimmed.
• One definition per number. What is counted, since when, from which source: one sentence under each indicator avoids arguments in meetings.
• The right chart. Bars to compare categories, a line for change over time, pie charts only for a few shares of a whole.
• The right permissions. An internal dashboard must be restricted to the relevant groups, like the map that goes with it.
• An owner. Someone must regularly check that the data is current and the numbers remain correct, and must be the person users contact when a figure looks wrong.

Think about performance too. Because elements are recalculated on every display, a heavy query on a large table slows down the whole dashboard. Prepare complex calculations in PostGIS views, index the filtered fields, and test the display with the real data volume before presenting the dashboard to its audience.

Usage examples

A few typical situations, to picture what each tool brings:

In each of these cases, the approach is the same: start from the question users ask, check that the data needed exists and is clean, choose the simplest tool that answers it, even if a more sophisticated one is available, then show a first version to the people concerned before enriching it. A dashboard is built in steps, not in one go, and the first version is rarely the last.

These examples illustrate the features described above; they are not customer references.

Which tool for which need

Comparing these tools with proprietary dashboards? The guide Lizmap vs ArcGIS Online covers that choice for local governments, and the guide What LizPack does explains what LizPack takes care of around Lizmap.

Frequently asked questions