2014 West Africa Ebola Response/UNMEER Routing Project
Jump to navigation
Jump to search
Tool Options
GPS Data
April 1 Meeting
April 1 meeting to summarize coordination to date
Summary of Robert's Data Collection Update
- Primary area involved Kambia in Sierra Leone
- Organizations involved: Red Cross, Mari Steps, UNMEER Fleet Manager, Drivers union of Kambia
- Currently some paper maps on office walls, drivers write travel times along routes they took.
- WHO Ebola surveillance team in Kambia uses OsmAnd for GPS tracks, 30 second interval
- Currently GPS traces are saved to Google Drive
- E-health soon to start Ebola surveillance project for vaccine trials, they are requested to use OSMAnd or OSMtracker to collect GPS traces as well
- Need SOP for OSMAnd installation (Robert created this)
- Need additional training materials for new OSMAnd users
- Port Loko (45 mins away) might join OSM GPS trace collection project
- Paper maps - NERC will scan and distribute to UNCS, HOT/OSM and NERC staff
- Each group to evaluate validity and usefulness as an emergency collection process
- 9000 phones were donated for Ebola
Improvements Needed
- Automate upload of GPS traces
- Travel time information added to road objects in OSM (average_speed=) based on time of year (OSM conditional tag for month)
- Development of isochrone scenario planning tool for Health Facilities, Markets, etc
- Ability to upload layers (shapefile?) of points for isochrone tool
- Select time frames for isochrone tool
- Export results (shapefile?) from isochrone tool
- Continuing collection of GPS traces to get info for all months and improve travel time info with more traces
- Consider average_speed reset after some time of no new updates
Discussion Highlights
- OSRM has an isochrone tool ([Libraries.io https://libraries.io/npm/osrm-isochrone])
- GraphHopper has a closed source isochrone tool (not clear if this is the closed source tool or not)
- osrm-isochrone would work for graphhopper, but again: this is a purely JS solution and imprecise as has to assume fixed speed of 70mph
- OpenTripPlanner has an isochrone tool, but not suitable for large areas
- OSM is more and more used for offline maps with PoIs, ebola surveillance, contact tracing.
- There are some good links here: Isochrone
- To properly detect an actual change, and distinguish it from a spurious value, would be good to have a record of recent values and calculate variance
- We could propose and use an average_speed tag for OSM, but an out of band process should generate the suggested tags and a manual import process to attach that info the OSM road segments.
- We should use a workflow that requires HOT/OSM person to review the generated average_speed info and approve adding it to the OSM live data.
- Paper map workflow is good for emergencies as a first, low technology way to gather speed related data
- Imapassable roads can not be automatically detected and will require someone to mark a spot "impassable" by hand in the phone app
- Additonal features like marking points for users should be considered a future update.
IRC Log
| 16:41:02 | <LuisHernando> | Hello friends |
| 16:45:16 | <BlakeGirardot_> | Hi LuisHernando |
| 16:46:35 | <BlakeGirardot_> | This is where we thought we would have the coordination meeting. |
| 16:55:54 | <pierzen|2> | Hi LuisHernando, meeting in 5 minutes :) |
| 17:00:05 | <pierzen|2> | UNMEER routing discussion should start soon. Waiting for more people to participate |
| 17:00:34 | <mkl> | hi I am here for the UNMEER routing discussion |
| 17:01:29 | <RyanLash> | I am here for the UNMEER routing discussion as well |
| 17:01:31 | <clairedelune> | I'm also here for the meeting |
| 17:01:43 | <mkl> | i just pinged Robert Colombo |
| 17:02:02 | <pierzen|2> | Hi all, I announced on the OSM irc to bring in more people. |
| 17:02:02 | <mkl> | and Peter from GraphHopper |
| 17:03:04 | <ptressel> | Likewise here for routing. |
| 17:03:57 | <robert_vram> | robert vram |
| 17:03:58 | <robert_vram> | i am here |
| 17:04:07 | <robert_vram> | Accessibility tool. |
| 17:04:33 | <LuisHernando> | hello |
| 17:04:34 | <LuisHernando> | Hola Tio |
| 17:05:40 | <pierzen|2> | Do we start the discussion now, or should we wait for more participants? |
| 17:05:58 | <robert_vram> | let s wait form kilen and the Graph hoper guys |
| 17:06:00 | <mkl> | LuisHernando: will Robert be joining? |
| 17:06:25 | <LuisHernando> | Hello everybody. I'm Luis hernando AGUILAR Information Officer in UNMEER (@luishernando) I used to work in OCHA and I participated in several emergencies (I see some well know nicks here) |
| 17:06:41 | <LuisHernando> | Robert is robert_vram |
| 17:06:48 | <ptressel> | Folks who arrive late can catch up on the log at logs.sahanafoundation.org/hotosm/2015-04-01.txt |
| 17:07:20 | <mkl> | ok great |
| 17:07:28 | <mkl> | I just pinged Peter again, hopefully he can join |
| 17:08:12 | <robert_vram> | thanks Pat |
| 17:08:59 | <mkl> | well let's get started, we have a critical mass. and i think the bigger questions are in data collection methods, tags, etc |
| 17:09:04 | <mkl> | so we can start there |
| 17:09:34 | <LuisHernando> | Lets go! :) |
| 17:09:58 | <BlakeGirardot_> | Can someone update on the status of data collection ? |
| 17:10:08 | <robert_vram> | i can |
| 17:10:23 | <robert_vram> | currently several activities taking place |
| 17:10:50 | <robert_vram> | focusing in Kambia in SLE as the pilto project site..we say pilto as stuffis going on there frmo different org |
| 17:11:23 | <robert_vram> | OMS paper maps have been hanged on several offices and we have requested the driver to write a log of where thy have been on a daily basis, adding info on Travel time for the route they tool |
| 17:11:46 | <pierzen|2> | Plus we started remote update of map for Kambia http://tasks.hotosm.org/project/971 |
| 17:11:49 | <robert_vram> | offices involve REd cross, Mari Steps , Fleed manager from UNMEER, Drivers usion of Kambia |
| 17:12:19 | <robert_vram> | WHO suveillance team in Kambia are starting to use OSM and to collect GPX traces with points every 30 secs |
| 17:12:45 | <robert_vram> | the GPx traces are now beeing placed in a google drive. the owner of the drive needs to give us approval to sahare it with other partners |
| 17:12:56 | <robert_vram> | E-health will implemlent a surveillance project for Vaccine trail there |
| 17:13:13 | <robert_vram> | we have requested that they install OSM And or tracker on their phones to get GPX traces |
| 17:13:37 | <robert_vram> | project still not started. curernt ly writting an OSP of how to install OSMAND fo rnew users and WHo logistical officer information |
| 17:13:44 | <robert_vram> | TRaining will be needed for those folks |
| 17:13:50 | <robert_vram> | Basically this = Kambia |
| 17:14:09 | <robert_vram> | i will have a conversation with WHo staff in port loko that would like to join OSM GPx tracing |
| 17:14:25 | <robert_vram> | similar to what kambia is doing 9 they are separated 45 mins only in any case |
| 17:14:37 | <robert_vram> | whay is expected wit the info |
| 17:14:43 | <robert_vram> | Paper maps will com to the NERC |
| 17:14:46 | <robert_vram> | NERC will scan them |
| 17:15:10 | <robert_vram> | maps and info on TT will be made avaailbel trouigh scanned copyes to UNCS, OSM and NERc staff |
| 17:15:44 | <robert_vram> | each unit will work on them and add them to thjeuir layer? why 3 projects? to allow each Agency to test and try and engage them if they see the cvalidity of having this as a corporate solution for emergecneis |
| 17:15:50 | <robert_vram> | What is the request on OSM and |
| 17:16:12 | <robert_vram> | having the info updated autiomatically in a server, folder, without the user needing to be involved |
| 17:16:41 | <mkl> | robert_vram: so the idea is that GPX files and scanned maps with notes would be uploaded and shared for processing by HOT? |
| 17:16:51 | <robert_vram> | OSM will need to figure out how to use this info and update the road layer with TT info that will chane as meny users will use the Application |
| 17:16:57 | <robert_vram> | expectur results |
| 17:17:08 | <robert_vram> | mikle you are right |
| 17:17:12 | <robert_vram> | let me continue briefly |
| 17:17:21 | <robert_vram> | Expected results |
| 17:17:28 | <robert_vram> | routable layer from OSM |
| 17:17:52 | <robert_vram> | developement of an isochrone tool for scpenario planning 9 location of HF, Markets, etc |
| 17:18:19 | <robert_vram> | the more the users run on the layer the more precise the information for each month season will be |
| 17:19:01 | <robert_vram> | with isochrones based on matrix of location s9 i would suggest layers of points instead to avoid another step) scenarios will be able to be snalyzed |
| 17:19:12 | <robert_vram> | how a road evolves on dry rainy season is veru sensible in Africa |
| 17:19:16 | <robert_vram> | Future outcomes |
| 17:19:50 | <robert_vram> | if a tool can be implemented on line , humanitaian comunity will be abel to use for their own calculations engagig tehm to be information provider not just for roads but for location (HF, CCC etc |
| 17:19:58 | <robert_vram> | i am done from my side |
| 17:20:26 | <BlakeGirardot_> | Perfect, thank you very much. I was behind in the info. |
| 17:20:52 | <clairedelune> | who is developping the isochrone tool? |
| 17:21:11 | <robert_vram> | how do we proceed? what do developers think on those? As far as i know ther is no isochrone tool beeing implemented yet |
| 17:21:19 | <ptressel> | OSRM has one. |
| 17:21:39 | <robert_vram> | grashoper? |
| 17:21:45 | <robert_vram> | graphhoper sorry |
| 17:22:19 | <BlakeGirardot_> | Are there offline requiremtns for the routing tool? |
| 17:22:29 | <mkl> | karussell has joined from graphhopper |
| 17:22:39 | <robert_vram> | would be good to have tehe capaity to upload layers of elements to make the calculations |
| 17:23:04 | <mkl> | my question is how we intend to process the GPX and scanned maps, in order to feed into data for the routing tool |
| 17:23:15 | <karussell> | there is a currently closed source of such an isochrone tool for graphhopper. also opentripplanner has one but not suitable for large areas |
| 17:23:23 | <robert_vram> | basically how do we get the data from the users on the field without neede hem to export manually the data and how this data on TT is added to the routable layer to be considered as an attribute? |
| 17:23:44 | <robert_vram> | mkl good question |
| 17:24:11 | <ptressel> | http://libraries.io/npm/osrm-isochrone |
| 17:24:15 | <robert_vram> | as we use OSMAND (or the proposed evolution of the tool) all gpx would be having the same structure |
| 17:25:29 | <robert_vram> | could OSM and or tracker be modified to do this? i klnow google mytracks can add al the data automatically to a Google drive |
| 17:25:37 | <robert_vram> | just questions |
| 17:26:12 | <BlakeGirardot_> | osmAnd supports plugins |
| 17:26:17 | <karussell> | ptressel: as it is a JS solution would work with nearly any router BUT this is a rather imprecise solution ... |
| 17:27:44 | <ptressel> | Looks like the main repo is hosted by mapbox: https://github.com/mapbox/osrm-isochrone |
| 17:28:14 | <robert_vram> | looks promissing |
| 17:28:59 | <ptressel> | I don't have my notes from the isochrone discussion some months ago :( but I did chat w/ someone from OSRM. They were very open to working on it some more. |
| 17:29:29 | <BlakeGirardot_> | and osmtracker is opensource so I don't see why either or both of them could not be made to support the auto uploading of the gps traces. |
| 17:29:50 | <mkl> | karussell: you had some options right? |
| 17:30:14 | <mkl> | even if closed source, if it works for the requirements, we can consider it |
| 17:30:17 | <robert_vram> | now how do we make it possible: add GPX through OSMAnd data, Addapt OSM App to automaticall update info, OSM TT availability ( selection of time frame) ...to later go to online website..using the layer a user would be able to update a point shapefile with codes of elements to map, select isochione calcualtion (select time frames ) export resulting |
| 17:30:17 | <robert_vram> | isochrone as shape for heyr own consumtion |
| 17:31:02 | <robert_vram> | people on he field are really using OSm guys... |
| 17:31:22 | <robert_vram> | and they are starting to see the benefit of ofline maps in their phones for POI, cuurveillance, contact tracing |
| 17:31:32 | <clairedelune> | :) |
| 17:32:35 | <robert_vram> | i spoke to IFRC colleague and he told me:" i would like to be ablo to define where i shiould train people to make sure i cover the compelte territorory of SLE and i leave some capacity here...but can not define where now because I do not know how long it takes me to go from opne place to the opther in this country , and do not know how it will be i |
| 17:32:35 | <robert_vram> | n the rainy season" |
| 17:33:22 | <robert_vram> | see...advoacy is key..dont expect to have EPi team knowing about mapps all the time..IM and GIS should priovide3 support..if we catch the eye of decision makers we will make it anywhere |
| 17:33:56 | <mkl> | ok so distinct questions to answer |
| 17:34:05 | <mkl> | 1) how to process collected field data into OSM |
| 17:34:18 | <mkl> | 2) how to present that data as an appropriate routing application |
| 17:34:37 | <mkl> | anyone have ideas for 1? |
| 17:34:38 | <karussell> | mkl: yes, see https://github.com/graphhopper/directions-api#isochrone-api |
| 17:34:58 | <mkl> | robert_vram: can you share collected sample data so far? |
| 17:35:30 | <robert_vram> | i can share some but not all the google drive |
| 17:35:31 | <karussell> | also the osrm-isochrone would work for graphhopper, but again: this is a purely JS solution and imprecise as has to assume fixed speed of 70mph (just look into the source) |
| 17:35:46 | <mkl> | karussell: that looks perfect actually. that is all from OSM data? someone would just need to write something against the API? |
| 17:35:48 | <ptressel> | That can be fixed. |
| 17:35:48 | <BlakeGirardot_> | mkl, I feel like you left out 0) how to automate the upload of the gpx traces unless that is not an issue for some reason |
| 17:36:04 | <karussell> | mkl: yes, from OSM |
| 17:36:28 | <mkl> | karussell: 70 mph fixed speed in osrm-isochrone? but not graphhopper api? |
| 17:36:32 | <karussell> | mkl: re 1) if it is traffic data it won't belong into OSM, IMO |
| 17:36:48 | <mkl> | what tags does it utilize to calculate road segment speed? |
| 17:36:56 | <ptressel> | There are some good links here: Isochrone |
| 17:37:03 | <mkl> | or how else is that data stored, if not in OSM tags? |
| 17:37:29 | <mkl> | BlakeGirardot_: not sure that's essential. robert_vram? there does at least need to be a workflow to upload, automated or not |
| 17:37:56 | <robert_vram> | it need to be an automated format |
| 17:37:57 | <karussell> | mkl: we created a solution for graphhopper which can calculate speed based on surface, highway, etc |
| 17:38:07 | <robert_vram> | paper maps are fast temporary solutions frmo peple from the field |
| 17:38:21 | <robert_vram> | but GPX traces shoudl be teh main tool |
| 17:39:05 | <robert_vram> | for places like kambia whern a solution need sto be provided taht was the way to engage non technical users (as it moslty happen in emergencies drivers do not have smartphones al lthe time,. specially in Africa) |
| 17:39:24 | <robert_vram> | karussell: i have seen models using that |
| 17:39:35 | <robert_vram> | now in an earthwuale or flood the speed is marked by other elements |
| 17:39:54 | <robert_vram> | thats the goal..a tool that it is updated by the user in a particular setting during a crisis |
| 17:40:09 | <mkl> | anything else karussell? we are not trying to collect traffic, but average speed on a road |
| 17:40:09 | <pierzen|2> | Talking of OSM, the first step is to define tags to add to the database. We need a tagging schema that routing tools can use. |
| 17:40:22 | <karussell> | robert_vram: yes, sure. It is not hard to adapt that speed calculation |
| 17:40:25 | <mkl> | that's going to be somewhat dynamic, main variable will be season |
| 17:40:43 | <robert_vram> | mkl |
| 17:40:49 | <karussell> | mkl: I see, then probably okay via the traffic tag? |
| 17:40:50 | <robert_vram> | season related |
| 17:41:01 | <BlakeGirardot_> | OSM supports that now as well mkl. I forget the name of the tagging schema but the (may - june) format |
| 17:41:05 | <robert_vram> | or even event related...think of an earthwuake |
| 17:41:16 | <robert_vram> | eartquake where roads are unpassable |
| 17:41:22 | <robert_vram> | let s get the users get us the data |
| 17:41:58 | <robert_vram> | help identify critiacal points on a road 9 aparticular damaged section, normally bridges) that could boos accessibility |
| 17:42:21 | <karussell> | Conditional restrictions |
| 17:42:30 | <BlakeGirardot_> | I envision the data on the road segment with the time period for each month of year generated automatically from the gpx traces. |
| 17:43:37 | <robert_vram> | blake automatically udpated |
| 17:43:45 | <BlakeGirardot_> | yes |
| 17:43:45 | <robert_vram> | let s say a user passes for the first time the road |
| 17:43:49 | <robert_vram> | the data is collected |
| 17:43:59 | <robert_vram> | now 10 users pass next day at differnet speeds |
| 17:44:12 | <robert_vram> | an average speed could be calcualted and update the precious one |
| 17:44:26 | <robert_vram> | let's face that we will not ghet the vehicle the user is using |
| 17:44:30 | <BlakeGirardot_> | That is what I envision. It just gets more accurate over time. |
| 17:44:52 | <robert_vram> | but after 100 uses and 100 collected points average speed (or median speed or any opther statitically chosen speed) cold be used and set |
| 17:44:58 | <robert_vram> | yep |
| 17:45:02 | <robert_vram> | accuracy is user dependent |
| 17:45:38 | <robert_vram> | that the trick to make it usefull for users: open, user related, udpated over time, adaptable to situation |
| 17:45:56 | <BlakeGirardot_> | I dont understand what "user related" measn |
| 17:46:04 | <clairedelune> | But in some places after a few years, another layer of gravel is added and then suddenly vehicles goes faster... would be good to be able to detect this kind of change as well in future. |
| 17:46:13 | <ptressel> | What about retaining other info besides the average speed, e.g. at least the variance? |
| 17:46:15 | <BlakeGirardot_> | I agree claire |
| 17:46:28 | <robert_vram> | user will be adding this info |
| 17:46:39 | <robert_vram> | a tag of last collected data would be usefull maybe |
| 17:46:46 | <robert_vram> | automatically |
| 17:46:53 | <robert_vram> | let's start step by step |
| 17:47:03 | <robert_vram> | the curent use is for emergencies where people move |
| 17:47:10 | <robert_vram> | roads are also being updated |
| 17:47:20 | <robert_vram> | and season makes road unpassable from one day to the other |
| 17:47:51 | <ptressel> | To properly detect an actual change, and distinguish it from a spurious value, would be good to have a record of recent values. |
| 17:47:58 | <robert_vram> | algoritm should be able to detect what is happenign if after 1 year of ntohaving a single poinjt collected in aroad set to 30km h , the new recorded speed is 100 |
| 17:48:09 | <mkl> | BlakeGirardot_: i think automatically updating speed from GPS Traces into OSM is problematic |
| 17:48:27 | <robert_vram> | but it is refining the alroith of average speed setting then..not the OSM tag aquisition for speed |
| 17:48:29 | <mkl> | there should be some intermediary step with involving human eyes |
| 17:49:05 | <robert_vram> | well paper maps could give you this if it takes part of the first assessmetn combined |
| 17:49:10 | <BlakeGirardot_> | I agree mkl when you say it that way. It would automatically generate the data and a human would review and actually attach to the road segment. |
| 17:49:20 | <robert_vram> | humans in emergencies do use the road but do not assess them |
| 17:49:26 | <karussell> | mkl: because of this problem I still suggest an external step for this automatically created speed suggestions |
| 17:50:05 | <robert_vram> | it is for emergencies specially |
| 17:50:16 | <mkl> | do we have an option for processes the GPX files into speed suggestions? |
| 17:50:27 | <BlakeGirardot_> | robert_vram: I understand what you are saying. You are right "impassable" is a challenged as there is no good automated way to detect impassable. |
| 17:50:31 | <robert_vram> | TT will not be usefull when humanitarians leave as we would not be able to control de users then |
| 17:51:07 | <robert_vram> | you are right Blake. |
| 17:51:41 | <BlakeGirardot_> | But there are also other data points you want the users to easily collect and report to HOT/OSM |
| 17:51:42 | <robert_vram> | starting poitns shopuld already making assumptions of speeds |
| 17:51:53 | <pierzen|2> | Impassable or difficult : we dont know if somebody was stopped two hours for some rest or because of obtacles on the road. |
| 17:51:57 | <robert_vram> | to latter have them updated by the OMS trcker users |
| 17:51:59 | <ptressel> | If you want to detect whether a new reading that is far from the previous average is statistically improbable, one needs some info about the distribution, not just the average. |
| 17:52:36 | <robert_vram> | repetitions of stoped people on a point during an emergency would mean obstacles |
| 17:52:49 | <robert_vram> | if one passes fast it might be road block o control |
| 17:52:50 | robert_vram RyanLash Richlv reidab rekth rorym_ rup3rt russss_ | |
| 17:52:55 | <clairedelune> | Impassable roads can be edited manually using this tag: Humanitarian OSM Tags/practicability |
| 17:53:02 | <ptressel> | Or, it might mean there's a rest stop there. |
| 17:53:16 | <robert_vram> | forget about user to edit on OSM. We are taklkkign about non GIS or mapping users |
| 17:53:28 | <BlakeGirardot_> | robert_vram: Would something like a button to mark things work? |
| 17:53:32 | <BlakeGirardot_> | So on their phone |
| 17:53:34 | <robert_vram> | 100% stoping at the same location? not reallywhile on the field |
| 17:53:35 | <BlakeGirardot_> | they are stopped |
| 17:53:43 | <BlakeGirardot_> | they hit a button to mark "blocked at this point" |
| 17:53:54 | <BlakeGirardot_> | or bridge/ford/etc |
| 17:53:55 | <robert_vram> | button could be. OSM and has one already to take pics and notes |
| 17:54:07 | <robert_vram> | osm tracked has that |
| 17:54:26 | <robert_vram> | but can t be sure that the surveillance team will click it |
| 17:54:32 | <pierzen|2> | osmtracker? |
| 17:54:38 | <robert_vram> | i can make sure he will not forget the phone for sure, specially contact tracers |
| 17:54:39 | <BlakeGirardot_> | Ok, cool, just wondering if we could get it down to like 10 big buttons for the 10 most commonly needed to record item would work for them. |
| 17:55:17 | <clairedelune> | robert_vram I meant that instead of using the speed tag, we could fill the practicability one instead by "clicking on a button" |
| 17:55:27 | <robert_vram> | Bklake OSm tracked does taht buit I think it does not have the offline capacity to have maps that user can see..Tis is another conideration...let s make user start to use maps |
| 17:56:16 | <robert_vram> | in y opinion: user should not be involved...whe the y see the added value of the data they were collecting, we can think of udpates |
| 17:56:40 | <pierzen|2> | It would be good that the various OSM products use the same basemap. This obf OSMand file could be shared by various applications. |
| 17:56:41 | <robert_vram> | contact tracers go to the field to trace possible infected people. Will not have time to continously update road status..it is nto theyr job.. |
| 17:57:23 | <BlakeGirardot_> | I see so two different user bases, professional drivers and people who drive as part of another professional activity. |
| 17:58:02 | <robert_vram> | professional drivers ....another story...paper maps for them |
| 17:58:18 | <robert_vram> | simple people..they drive..and we can not make use the phone while driving legally |
| 17:59:45 | <clairedelune> | when a road is really impassable, usually they are stopped (and could maybe click on a button just in that case) but I fully agree everything should be as much as possible automatized |
| 18:00:10 | <robert_vram> | if it is a corporate solution that would do..but drivers are not my concern...are the Smarthopne users.. |
| 18:00:24 | <robert_vram> | lets thing big..9000 phones were donated for Ebola |
| 18:00:33 | <robert_vram> | how many ogf them are they using GPX tracers ? |
| 18:00:35 | <robert_vram> | no many? |
| 18:00:35 | <BlakeGirardot_> | robert_vram: I got lost on the user base then. So automated gpx traces only from non-professional drivers? |
| 18:00:55 | <robert_vram> | those who use GPX what do they do with ti? they use it internally |
| 18:01:06 | <robert_vram> | how many send info to OMS or other mapping agencies? |
| 18:01:22 | <clairedelune> | very few |
| 18:01:27 | <robert_vram> | Blake gpx trace sfor any user diregarding who they are |
| 18:01:51 | <robert_vram> | if a driver gets a smartphones let s make sure he knows how to clik on/ffGPX tracing when he leaves and come back |
| 18:02:21 | <BlakeGirardot_> | Ok, so some drives might end up with smartphones for automatic gpx collection |
| 18:02:34 | <robert_vram> | yep |
| 18:02:40 | <robert_vram> | we could push for that. |
| 18:03:03 | <robert_vram> | As mentioend i am writting an SoP for installing OSM and..but even for me it was not eay in the begining |
| 18:03:31 | <robert_vram> | so ...similar to crisis signal app install , say how you want to share..done |
| 18:03:42 | <robert_vram> | will just need hat driver starts the GPS |
| 18:03:51 | <BlakeGirardot_> | Do you have a preference yet for OsmAnd or OsmTracker personally? |
| 18:04:17 | <robert_vram> | OSMAnd as you can have a refernce map and see your trace..important for some activities |
| 18:04:31 | <robert_vram> | but i was willing to hear from other developers about their recomendation |
| 18:04:31 | <pierzen|2> | Other solution for trace collection is these cheap gsp sniffers that can trace for a complete week. |
| 18:04:44 | <robert_vram> | smartphones |
| 18:04:49 | <robert_vram> | no one loose them |
| 18:04:55 | <BlakeGirardot_> | Ok, so there is some functionality in OSMAnd you use that tracker does not have. Got it. |
| 18:05:05 | <robert_vram> | well i did but everyone takes care of their staff if they see a use for the device |
| 18:05:19 | <robert_vram> | snifer would do for Coca cola and logistics tracing..another story |
| 18:06:03 | <BlakeGirardot_> | Ya, I had that idea too pierzen|2 Tiny tracker, solar powered, auto connect to wifi and dump traces every 24 hours or when in wifi range. Just mount it to the truck, done. configure via bluetooth smartphone app once. |
| 18:06:20 | <robert_vram> | Blake go didea |
| 18:06:24 | <robert_vram> | but anotehr device |
| 18:06:28 | <BlakeGirardot_> | ya :( |
| 18:06:36 | <BlakeGirardot_> | That probably doesn't exist yet |
| 18:06:39 | <robert_vram> | no donations for them yet....maybe in 10 years...implement the same on the smartphone please |
| 18:06:51 | <robert_vram> | this is exaclty what is needed |
| 18:07:01 | <robert_vram> | second stage woul be how this info is treated at OSM |
| 18:07:05 | <robert_vram> | routable layer |
| 18:08:06 | <ptressel> | GPS tracking devices for vehicles certainly do exist -- they are used by shared car agencies, for monitoring driver behavior by insurance companies... |
| 18:09:10 | ptressel pierzen|2 pgiraud Plesioth pnorman ProN00b | |
| 18:09:28 | <LuisHernando> | but I think that is not realistic to have those devices in that area |
| 18:09:42 | <ptressel> | Do a web search for "vehicle mounted gps tracking". |
| 18:09:55 | <LuisHernando> | even the biggest groups have not that |
| 18:10:17 | <clairedelune> | some humanitarian organizations (UN) have got tracking chips on their vehicles as well but I can't remember the frequency of the record |
| 18:10:22 | <BlakeGirardot_> | ptressel: Yes, but I was thinking more a "config, stick on truck, forget about it" no wiring or microsd cards etc. And ya, just not part of this unfortunately at the moment. |
| 18:10:31 | <LuisHernando> | I understand that there are some agriculture companies that have some, but is not in the most prioritized area |
| 18:10:51 | <BlakeGirardot_> | So the data in OSM |
| 18:11:07 | <robert_vram> | UN? |
| 18:11:12 | <LuisHernando> | clairedelune sadly for us in UN is not an standard |
| 18:11:17 | <robert_vram> | well tell me who has thm in Ebola and i ll contact them |
| 18:11:19 | <ptressel> | BlakeGirardot_, Yes, there are companies that provide that service, but commercial. |
| 18:11:28 | <robert_vram> | so Far I have seen Barret rado but GPS in use? nonononononon |
| 18:11:39 | <BlakeGirardot_> | I think we can cover it with either the existing tags or maybe we propose an average_speed=* tag |
| 18:11:41 | <robert_vram> | fleet management shoudl be corporate but it is not there |
| 18:11:45 | <robert_vram> | not now. |
| 18:11:45 | <LuisHernando> | the MOSS say that a GPS device must be in the car, but the truth is that the devices is in the car, but turned it off into a box |
| 18:12:20 | <robert_vram> | yep |
| 18:12:26 | <robert_vram> | smartphones are always in pockets |
| 18:12:31 | <clairedelune> | I know this from some other countries, will get back to you later if I get confirmation that some have it in the area. |
| 18:12:31 | <robert_vram> | use what the user has |
| 18:12:38 | <clairedelune> | agreed |
| 18:12:45 | <robert_vram> | in any case we look for other users also |
| 18:12:46 | <LuisHernando> | (I tried to involve to UNDSS on this and there are some nice examples but in Colombia, not here ) |
| 18:12:57 | <robert_vram> | GPX traces , OSM TT update, Isochrone calcualtion |
| 18:13:32 | <BlakeGirardot_> | I just don't see any issue but programmer time |
| 18:14:34 | <robert_vram> | let s try then |
| 18:14:48 | <robert_vram> | what about OSM layer update? |
| 18:15:00 | <robert_vram> | mkl wold you be wiling to reeive the GPx traces and the paper scanned maps? |
| 18:16:24 | <BlakeGirardot_> | I can't speak for mkl, but I would be, understanding that the paper maps are just a temporary solution and that we need programmer time to process the gpx traces. |
| 18:17:19 | <BlakeGirardot_> | Then we would have to develop a workflow to attach the processed gpx trace data to roads in OSM |
| 18:17:24 | <robert_vram> | ok |
| 18:17:25 | <BlakeGirardot_> | that is what we can't fully automate. |
| 18:17:36 | <robert_vram> | i have sent once mroe the request to sahre GPX traces |
| 18:17:53 | <BlakeGirardot_> | I have a sample one you(?) sent |
| 18:17:59 | <ptressel> | Do the traces have timestamps on individual points? |
| 18:18:01 | <BlakeGirardot_> | Someone sent a sample. |
| 18:18:18 | <robert_vram> | i have a GPX trace yes |
| 18:18:25 | <robert_vram> | from kambuia |
| 18:18:44 | <BlakeGirardot_> | ptressel: The one I saw did and it had speed info attached, but I was told we are better off calculating speed. |
| 18:19:03 | <karussell> | BlakeGirardpt_: that is automatable via map matching: https://github.com/graphhopper/map-matching |
| 18:19:10 | <ptressel> | Ok, good. |
| 18:19:43 | <robert_vram> | https://www.filepicker.io/api/file/4SGsds0ZSIe4lYBOlPAl |
| 18:19:56 | <BlakeGirardot_> | karussell: I thought the main issue with automated into OSM was philosophical really. We use a tool like that to generate a .osm file, someone opens, reviews, and merges by hand, just because. |
| 18:20:34 | <robert_vram> | https://www.filepicker.io/api/file/3xOpJ3OeQfacwtBXGiDl |
| 18:20:35 | <BlakeGirardot_> | just because we want a human to look at it |
| 18:20:40 | <robert_vram> | freetown kambioa and Kambia Freetown |
| 18:20:49 | <robert_vram> | road is asphalt |
| 18:21:22 | <robert_vram> | humans make errors |
| 18:21:26 | <robert_vram> | 100 humans make a trend |
| 18:21:26 | <BlakeGirardot_> | And we might have to adjust/split/combine/otherwise clean up ways. |
| 18:21:36 | <Zverik> | As for car gpx trackers, I've testen one that's mounted with a magnet on a roof, and in comparison with a regular Garmin etrex traces were quite poor in quality |
| 18:22:59 | <ptressel> | I hear it's better for accuracy to mount them *under* the vehicle. Don't know the reason. |
| 18:23:01 | <BlakeGirardot_> | Zverik: Inteseting. I would expect a mag mount to be ..... problematic |
| 18:23:31 | <BlakeGirardot_> | I have to run for 15 mins. But I can catch up when I get back |
| 18:23:39 | <Zverik> | globalsat bu-353. Magnet worked excellent, because my car is made of steel |
| 18:23:42 | <BlakeGirardot_> | if I don't see you leave, I really want to thank everyone for attending |
| 18:23:43 | <ptressel> | GPS signals are received from satellite -- signal not affected by magne3t. |
| 18:23:44 | <robert_vram> | must keep back on my Accra stuff guys |
| 18:24:12 | <robert_vram> | will be available in any case...and in case i get diconeected for to long [email protected] do the trick |
| 18:24:29 | <BlakeGirardot_> | I think we can do this really. Programmer time. That is the resource is short supply. |
| 18:24:39 | <ptressel> | :D |
| 18:25:12 | <pierzen|2> | Thanks robert_vram for this presentation. We can follow by email and even bring interested people in the Skype discusion group. |
| 18:25:44 | <ptressel> | http://codeforall.org/ |
| 18:25:45 | <robert_vram> | ta=lk soon! teka care! |
| 18:25:57 | <pierzen|2> | same 2 you |
| 18:26:10 | <ptressel> | There are sources of programmers wanting to do humanitarian work. |
| 18:27:25 | <pierzen|2> | thanks ptressel |
| 18:27:48 | <ptressel> | That's only one organization -- there are a lot. |
| 18:28:19 | <clairedelune> | Thank you all |
| 18:28:50 | <clairedelune> | pierzen|2 please include me in the skype group |
| 18:29:13 | <pierzen|2> | ok |
| 18:29:14 | <mkl> | I wasn't able to follow the entire time, but will look for the notes and look through the logs |
| 18:29:22 | <clairedelune> | thx |
| 18:29:29 | <mkl> | and help however i can to move things forward |
| 18:29:49 | <ptressel> | There was also a chat yesterday re tagging. |
| 18:31:33 | <pierzen|2> | We received a link for UAV imagery if you want to look. Just a small zone available presently in Port-Villa |
| 18:31:43 | <pierzen|2> | zoom at osm:#map=21/-17.7458823/168.3210853 |
| 18:31:48 | <pierzen|2> | JOSM link |
| 21:12:58 | <BlakeGirardot_> | skorasaurus: ping |