GeoAnalytics Travel Area drive times vs. Google maps

[chasafd]

I have a question about GeoAnalytics drive times (travel areas) vs. Google Maps/Earth.  In the first example, Point A is Fire Station 4 (circle) and Point B is a home address (triangle).  Google maps says it takes 4 minutes (driving the speed limit I assume)

But with GeoAnalytics the 4 minute Travel Time (gray poly) shows it comes up short (also driving the speed limit I assume)

I have to go out to 6 minutes to reach the house:

Example 2: here Point A is Fire Station 3 and Point B is the same house, Google maps says 7 minutes either way you go:

I have to go out to 8 minutes to reach the house

I realize the second example is close enough, but the first example is off by a lot (4 min vs. 6 min).  If you were trying to determine where to put another fire station somewhere in town, this may not be accurate enough.  And I understand that emergency vehicles travel faster than cars and drive times need to be adjusted by a small factor, but in my examples both use speed limits (I assume) so it's a fair comparison.

Example 3: The drive time from Austin to San Angelo, TX is 3.5 hours based on Google maps and my personal experiences (driven it many times).  The GeoAnalytics map with cost = 3.5 Hours comes up roughly 25 % short.

Is there a config setting I can adjust to make this more accurate?  Thanks, I appreciate any help you can provide.

Chas Randal

Austin, TX

Note: These examples were generated with QlikView 12.2 Nov 2017 SR6 and GA June 2018

[Patric_Nordstrom]

Hi Chas,

The route calculations are not exact science I'm afraid. They are based on limited data and calibrated to be reasonably accurate for most cases. Google probably have the best data in business since they have the resources and users to collect data efficiently. Looking at results from different routing suppliers the results sometimes differ quite a lot. For instance for the example route you had from Austin to San Angelo, Google had 3h 28min and Here (at https://wego.here.com) 4h 4min. We do not have access to traffic data so calibrating has the effect that certain road types and road features (like road lights) are assigned speeds and delays to mimic real traffic speeds. Data is improving over time as is the calibration of various factors so hopefully we will get there.

Thanks for a good investigation of some routing use cases.

Thanks,

Patric