6. • QlikMaps adds sophisticated, interactive mapping and
location analysis to your Qlik dashboards.
• Plot points and/or territories: See where your assets,
sales territories, and target markets are located.
• Use with or without an existing GIS system.
• Works in any browser, including mobile/tablet devices.
7.
8. AGENDA
1. Why Maps Help Your Analysis
2. Creating Decision-Based Maps
• Data Profiling
• Custom Pop-Ups
• Large Data Sets
• Data Prep Considerations
• Write Backs
3. Questions
42. • Simple architecture: No extra server required
• Simple compliance: All 3rd party licensing included
• Simple development: Identical to other Qlik sheet objects
• Simple UI: No end user training required
• Simple deployment: Qlik extensions are easy to install
• Simple pricing: Priced by environment or tokens
• External Data: Reach outside with Qlik data model
• Full functionality from all devices
- Analytics8’s heritage is in big data and analytics
- Our approach to product development is different than others- we develop our products with the BI consultant in mind, based upon our years of experience in the field. It’s why our products are so easy to install and implement.
- We developed QlikMaps specifically for Qlik from the ground-up- it was not built in an ivory tower and then forced into Qlik. QlikMaps was born from customer requirements- when our customers needed more robust mapping capabilities, we built QM.
Give a basic overview of what QlikMaps is.
This slide sets the groundwork for the product we are talking about. I expect this to be very quick and push questions about it to the end.
Overview of what you’re going to talk about -
Wages Dashboard
This dashboard walks through how adding a map adds value to a dashboard.
Dashboard:
This sheet has no map. States are compared based on Wages by State but without geographical reference the it’s difficult to tell if there is any bleed over from State to State.
Native Map:
This sheet has replaced a tree map from the Dashboard sheet with a native Sense map. The map allows you to see the relationship between the states in a way you can’t with a tree map. Even though value has been added to the map, the features available in the Native Sense Map still allow for ambiguity in the data due to lack of finer controls of color. Also, KML files were necessary to build the map. KML files must be found as they are not provided by Qlik. Furthermore, finding KML files online does not guarantee quality or necessarily provide the license to use the KML file for business purposes.
Data Profiling:
This happens outside of the Wages Dashboard, but the transition is fairly seamless. In the Native Map sheet we spend some time talking about obtaining the KML files. Now we can show how simple it is to use our Global Boundary File Library. We ‘download’ a copy of the State.qvd from QlikMaps.com/download and join it to the facts table using the data profiler.
QlikMaps:
Once we’ve shown how easy it is to load in our shape files, we navigate to the QlikMaps sheet where we show a map using QlikMaps. This map is primarily intended to show feature parity with the Sense Native map. However, we make it a point to show how we have finer control over colors, pop ups, base maps and navigation than Native Sense maps.
Full Page:
In this sheet we blow up the QlikMaps object so that it consumes over half the page. We have moved a number of the navigation elements and State Comparison objects to the pop ups. This shows that with a properly build map that leverages QlikMaps popups, it’s easier to delegate more space screen releaste to the map object where most of State comparison is found.
You’ll have just talked about large data sets at the end of the demo – and then slightly recap – and then go into the next question: How many points/shapes can I put on a map.
We get this question a lot and the truth is, it’s hard to answer. It’s hard to answer because it depends.
We could ask the same question about a straight table. I’m sure there is some theoretical limit to the number of rows a straight table can handle, but we don’t hear about. We don’t hear about it because the usability of the table becomes overwhelming long before we hit the number of rows it can handle.
If a user asked you to build a straight table with 100,000 rows in it, you would likely ask them some qualifying questions to help manage the usability of the application.
How is this data rolled up?
Will users be making selections before seeing the table?
Is there a certain percentage that need to be seen while others must be drilled in to for better visibility?
We ask users the same questions with maps. The next couple of slides show you some examples.
* The following slides are based off of screen shots from an app. This was done because the app is very large and performs better through Access Point on the server. Since I don’t trust the reliability of the internet at Qonnections for demo purposes, I pulled of screen shots to make this easier to work through. A video of navigating through this app is available at QlikMaps.com under the Webinar ‘Maps that Matter’
Above is 3700+ points displayed over a single state. 3700 points is not a lot of points, but as you can see, the points sit on top of each other making it difficult to pull useful information out of the map.
If you were displaying this map, you might be trying to answer the question “Where is the highest density of bridges?” Looking at the map, there are 2-3 different locations that seem to have a higher density of points. I’m willing to bet that your eyes are simply looking for the messiest area and attributing the highest density value to that area. Think about that, you are attributing the highest value to the parts of the map you least can make sense of.
Let’s look at the same data set using a heatmap.
Suddenly, the story changes. The places we previously thought of as hotspots aren’t as hot as we thought they were. In fact, the highest density bridges are actually tucked away in a corner I bet a lot of you missed.
Let’s look at the two together
Even Heat Maps have their limits. When spread over a large area, medium density areas tend to get lost.
Approximating points is a much simpler method of reducing the number of points to a visually usable number, without losing an fidelity.
In the example above, we have rolled up 480,000 + bridges to 2300 bridges. We did this by chopping of the last couple of digits of the lat/long during the load process and aggregating the points. Notice that we can clearly identify problem areas that need to be investigated.
You’re probably thinking we lost fidelity, but we have not. Think of each point as a territory. In our data model, each ‘Point Territory’ is linked to every point below it, simply waiting to drilled in to.
Here we can see that we’ve selected a single type of bridge and found an area that has an inordinate number of bridges in disrepair.
Once lassoed, we are no longer looking a approximated points, but the actual Lat/Long of that point.
Each point has a pop that provides detail about the particular bridge that point represents. Double clicking takes us to Street View where can see damage to the bridge.
An alternative to points approximation is using a geographical area to roll points up to.
In the example above, 480,000 bridges are being represented by 24,500 zip codes. This makes for an impressive looking map. However, very important information is being lost.
Not all Zip codes are created equal. Some are very large while others very small. When we use QlikView to parse out the Zip Codes with the highest density of bridges, we get this.
When we use QlikView to parse out the Zip Codes with the highest density of bridges, we get this. Notice how difficult it is to see these zip codes. This means that the areas with the highest density of bridges are likely never getting seen on map with all the zip codes.
Binning makes all areas equal. This means that small Zip Codes aren’t lost but the map remains visually pleasing.
When the same selections are made, the areas with the highest density of bridges are more visable.
More examples and details on a webinar – Maps That Matter – go to analytics8.com/qonnections and scroll to the bottom to find a link to that webinar.
- We developed QlikMaps specifically for Qlik from the ground-up- it was not built in an ivory tower and then forced into Qlik. QlikMaps was born from customer requirements- when our customers needed more robust mapping capabilities, we built QM.