For years we have used ArcGIS’ Network Analyst to compute network distances and route-based service areas. In the past we found open source options to be lacking. We experimented, for example, with pgRouting in 2014 and our rough assessment was that it did not provide an adequate alternative. Details are in our blog post on the topic.
A lot of time has passed since that initial assessment, so in a recent tweet, I asked the Free and Open Source for Geography (FOSS4G) crowd whether they have tools they use and are happy with. There were some great options for investigation — though I warn you that this post will be somewhat unsatisfying given that we have not tested the options, mostly I’m providing a list for reference.
If you have used any of these tools please let me know in a comment or by e-mail. I would be particularly interested in any comparisons like the one we did comparing ArcGIS and Google Maps API.
My initial query
One of the biggest reasons we still have ArcGIS licenses is for Network Analysis (drive times, service areas etc). Does anyone have #foss4g tools for this they like? Last time we tried pgRouting (yrs ago) it didn’t feel fully formed yet, #gischat
— Zev Ross (@zevross) January 28, 2019
Tweet response summary
- pgRouting has come a long way!
- QGIS new processing tools for Service Area and Shortest Path in 3.0
- In R, Robin Lovelace has provided the stplanr package
- ProjectOSRM (Planet-wide OpenstreetMap based routing)
- GraphHopper (got two votes)
- OpenTripPlanner
- openrouteservice.org
- osmnx, networkX and PySAL’s Spaghetti
- pandana (which stands for “pandas for network analysis”) in #python. It was developed by @fscottfoti @Waddell_Paul et al.
- Hqgis is a python based plugin for QGIS that offers access to the HERE API and combines different traffic/routing/geocoding actions in one plugin.
Tweet responses
You also have @ProjectOSRM and @graphhopper
Anyway, @pgrouting has evolved a lot // @VickyVvergara
Have you tried some @qgis? Maybe some network analysis plugin can do the work.
— María Arias de Reyna (@delawen) January 28, 2019
you may want to check out R pkg stplanr, Cc: @robinlovelace
— Edzer Pebesma (@edzerpebesma) January 28, 2019
then I think the main opensource options would be pgRouting and Valhalla
— Keith Jenkins (@kgjenkins) January 28, 2019
Not full replacement, but there's a start in a combination of @gboeing osmnx, networkX, and PySAL's Spaghetti
— Dani Arribas-Bel (@darribas) January 29, 2019
Not sure, but @ors_news may be worth checking out.
— Daniel Possenriede (@dpprdan) January 29, 2019
@pgrouting is solid and mature. It is a big open source project with a lot of features. @pgrouting community is totally awesome
— Cayetano Benavent V. (@cayetanobv) January 29, 2019
I’ve used GraphHopper, OpenTripPlanner and a bit of pgRouting. All work for my needs very well.
— Topi Tjukanov (@tjukanov) January 28, 2019
.@robinlovelace and @bikesRdata are also doing a tremendous work in developing tools for route network analysis in #rstats https://t.co/5pFLt0cnue
— Urban Demographics🚡 (@UrbanDemog) January 29, 2019
Looked at this for a project in grad school – the road network you use can have more impact than the algorithm or platform. Check out the difference between @ORS (interior red fill) and Here (blue exterior ring). ORS better captures the ‘can’t get there from here’ of Vermont. pic.twitter.com/jOYC2kM252
— David Fox (@ReynardSubtil) January 29, 2019
Interested in open source software with a command-line interface (CLI) for #reproducible analysis of route networks? Lots out there, e.g.:
– #rstats: intro https://t.co/kMd7h817SI + stplanr + dodgr 📦
– #python 📦: #osmnx
– @GRASSGIS/@qgis
– Thread: https://t.co/P6MRjWf49s pic.twitter.com/fHBPwnNpuJ— Robin Lovelace (@robinlovelace) January 28, 2019
Zev, don't forget pandana (which stands for `pandas for network analysis`) in #python. It was developed by @fscottfoti @Waddell_Paul et al. I believe @gboeing mentioned it in his previous tweet
— Urban Demographics🚡 (@UrbanDemog) January 29, 2019
also new from @ricckli 4 #QGIS –https://t.co/U4Mi0utDG5
— Geospex (@Geospex) January 29, 2019