How long does it take for new data to appear on the website?

We are working very hard to process new data and make it viewable on the map as quickly as possible. Scaling and adding new features to the site makes this more difficult but this is a top priority.

Currently the database of cell phone tower locations is updated about once every week or every two weeks.  Its done manually and thats painful.  The cellular heatmaps are updated much less frequently.  It represents a bigger challenge as the amount of processing power required to create the maps is non trivial and there are many kinks to work out.

The good news: We have technical challenges to work out but have made good progress.  We hope over the next few weeks to massively decrease the time it takes new cell phone towers to appear on the map with the not to distant goal of making this daily.

Heatmaps, by nature, are more difficult as every new point that comes in would require recalculation of our averages and thus each image tile needs to be re-processed. Each zoom has 2^n image tiles, so as you get past a zoom level of around 10, thats already a million images… We’re working on it though :)

Also, the stats on our homepage are updated every 60 minutes!

67 Responses to How long does it take for new data to appear on the website?

  1. At ITAPACI-GO BR, we have one tower of TIM and we got 2G.

  2. Scott Myers says:

    I have a custom ROM. In the network it reads Velocity instead of Verizon. Will this show in you data as a new network? Can you adjust for this? I drive all around my county and would like to have data for you guys, but would be useless in my opinion if it doesn’t have the right network data.

    • Sam says:

      Hi Scott,
      Thanks for letting us know! Looking at the data it is registering as a new network. There are a few users all recording their network as Velocity, and they are all in the US. It seems likely that all of these users are actually on Verizon.

      Our app records two pieces of data: the network name reported by the device and the network id. The network name is being given as ‘Velocity’, and the network id that is sent to us is invalid.
      What we will hopefully do soon is analyse the invalid network ids and assign them to a network manually. At the moment we are not sure how easy this will be to do for all the data we collect so I can’t give you a definite answer.

      Thanks again for using the app, and taking the time to report the problem.

      • Scott Myers says:

        The Velocity ROM is for the LG Ally only and the Ally is on Verizon only (I could be wrong). So I would think it would be safe to say to have anything Velocity default over to Verizon network. Also do you guys have a IRC chat room for ppl to join your team to maybe help out?

  3. Jared Martin says:

    Any chance of having the map have timestamps of when the cell towers info and when the heat maps were last updated?

  4. Dhirendra Kumar says:

    I seek your’ response

  5. Pierre Fagrell says:

    This was several months ago, how is the update rate of heatmaps and new towers now?

    • Sam says:

      Hi Pierre,
      We ran a heat map update on the 25th April. Currently heat maps are being updated roughly once a month, and tower locations every 1-2 weeks.

      • Pierre Fagrell says:

        It sounds good, but doesn’t seem to work in reality.

        I have now been using the app since April 26 and there should have been several updates of towers and atleast one update of heatmaps.

        However, there is NO tower on the map in my area, and no heatmaps.

  6. Hugo says:

    Are the towers from one company use by others company.
    Otherwise I can´t explain why I have a full reception signal and no tower of my company appears on the map
    Pilar- Argentina

    • James says:

      We don’t have every tower mapped, so it is possible your provider does have their own tower but it’s not appearing on the map. Using the app will help us discover towers and map them.

      It is also the case that carriers do share towers often. There are some carriers (Mobile Virtual Network Operators – MVNOs) who don’t own any wireless spectrum and may or may not have any towers. These carriers run by striking deals with operators that have infrastructure – either leasing or reselling capacity on other networks. But I’m not sure any are operating in Argentina.

      http://en.wikipedia.org/wiki/Mobile_virtual_network_operator

  7. jon says:

    Since I’ve tried repeatably over several months and never seen any signal strengths reported for my home, I have to conclude that this software doesn’t work (or is some type of scam). I deleted it from my phone.

    • Sam says:

      Hi Jon,
      Many apologies this comment was missed. Signal strengths you’ve submitted should definitely be on the map by now. If you send me approximate location (street + zip) I can check if we’ve received any data, and why it’s not displaying. My email is sam[at] this domain.
      Sorry again for the delayed reply.

  8. evert says:

    Working good! :-)

    I see the data supplied by my phone gets processed after a week or 2, and shows up on the map. Good job!

  9. Ssor says:

    When will the next update to the maps be? It has been over a month since the last update.

    • Sam says:

      Apologies for the delay. Our current database is failing when we try to run an update – we are currently in the process of moving over to a new one. We should hopefully be able to move up to more frequent updates within the next 1-2 months.
      Thanks for sticking with us!

  10. Bruce says:

    I just loaded the app today. I have Alltel that I don’t see a map for. I know it is a smaller carrier now, since Verizon bought most of the company. Will it have a separate map, or will it just show on an overall map?

  11. Hello,
    I’m not a user of OpenSignalMaps (I don’t have an Android device :/), but I’m really interested in the data.

    Is the data publicly available anywhere, in a format simple to download? I’d really appreciate if at least the tower data is made available :)

    Thanks,
    David

    • Sam says:

      Hi David,
      We are still in beta and we are still working on our public data. The raw data will be much too large to download – we’ll are looking at averaging the data and/or providing an API.
      Best regards,
      Sam

  12. Berthold Horn says:

    In our area a few antennas report latitude and longitude directly which can be read out on Android
    by ‘calling’ *#*#INFO#*#* and selecting “Phone Information” (The information is in hexadecimal
    and also needs to be divided by 14400 to convert to degrees). The local ones at least are
    quite accurate. Have you considered using these? (Or are they not always accurate, and
    not supplied by say Verizon Wirelsss e.g. for some reason, while U.S. Cellular seems to provide
    this information).

    • Sam says:

      Hi Berthold,
      We looked in to this many months ago. In our testing it seemed quite unreliable – there were times when we would get nothing back, or a lat/long that was very inaccurate. However it looks like for you it is providing an accurate location. We might start recording the data, then we can run some tests on it’s accuracy and see if there are patterns.
      Thanks for your feedback. It’s very useful!

  13. Berthold Horn says:

    Locally at least it appears that different antennas on the same tower have different CIDs (SSIDs).
    This obviously happens when different networks (e.g. Verizon wireless and U.S. Cellular)
    share the same tower, but also, antenna sets pointing in different directions (typically
    three different directions) are driven by different hardware and have their own “ID”.
    In OSM these show up in different places on the map because they are “seen” by
    cell phones in different locations (which introduces a different bias into the estimate).
    If you could identify which CIDs ‘belong together’ you could greatly improve accuracy.
    I am not sure how you would do this, although sometimes the CIDs are consecutive numbers.

    • Sam says:

      Hi Berthold,

      Yes that would be very nice. It’s hard to do, because of the inaccuracy in tower locations, and where you would draw the line to display towers separately. As our data gets more accurate and comprehensive we may decide to group antennas that are within a certain area – e.g. under 20m apart.

  14. Berthold Horn says:

    ‘Nanocells’ present a challenge because the cellphone connects locally to it, but
    the signal goes off over the internet. It ‘hops on’ the celullar network at a distant
    location. That place may be a hundred miles away (in my case), so the coordinate
    information collected is not useful, and potentially disruptive (oddly, the nanocell
    has to have a GPS receiver, so IT knows exactly where it is). Do you have a list
    of such places so you can ignore reports about them?

    • Sam says:

      Hi Berthold,

      Each network uses different identifies with there femtocells.
      In the US AT&T change the network name, so when your phone is connected you see ‘AT&T Microcell’ where you would usually just see ‘AT&T’. This is easy to filter out.
      In the UK Vodafone don’t change the network name – it stays as ‘vodafone UK’ whether you are connected to the SureSignal femtocell or a normal tower. It may be that there are other ways we can identify femtocells for Vodafone – but we haven’t been able to test. If anyone has a Vodafone SureSignal please send me an email at sam@ and I might ask you for a favour!
      What network are you on? If you have a femtocell please let me know!

  15. Steven New says:

    The different cell tower 3G or 2G indicators are wrong for where I live. In morpeth (ne61), UK there is an O2 network cell tower at

    http://opensignalmaps.com/index.php?lat=55.1781&lng=-1.7166&initZoom=13&isHeatMap=0&netwkID%5B%5D=23410&netwkType%5B%5D=3G

    which is NOT a 3G tower. The town of Morpeth is NOT covered by O2 3G at all. Also, The town of Alnwick (NE66), UK IS covered by O2 3G coverage, this tower is incorrectly labelled as 2G only

    http://opensignalmaps.com/index.php?lat=55.4056&lng=-1.6796&initZoom=14&isHeatMap=0&netwkID%5B%5D=23410&netwkType%5B%5D=2G

    Steven

    • Sam says:

      Hi Steven,

      Looking at the heatmap it looks like there was a very small O2 reading in that area. It’s likely to be caused by a phone that is recording it’s location incorrectly. For a variety of reasons sometimes a phone can record a location that is far from accurate. So it’s likely that this O2 user was far from Northgate Hospital.

      The second tower is being recorded as 2G because we do not have any user recording 3G signal from this tower. Our maps are currently a few weeks out of date, so it may be that we have correct data, and in the next map refresh it will be correctly displayed.

      • Steven new says:

        Thanks for the reply. Hopefully once the iPhone app comes out I can start feeding back some more realistic data in this area! Any estimates as to how long it will be for the iPhone app?

        • Sam says:

          Hi Steven,
          The iPhone app is a few months away from development. However unfortunately the iPhone software does not allow apps to access signal information. So our iPhone app will just be a viewer for our maps, as well as a 3G speed test.

  16. Rolf says:

    Hi, I want to ask why the application detects the position of the GPS module only when I switch on the “map”? When running in the background, leaving the GPS off. How then has to find a heatmap??

  17. Steve says:

    I live in the boonies and can barely get a signal at my house. I think this site is great, I was glad to find it. A question: There is an iPhone app called “signal” which displays signal strength and the location of nearby towers. So are you sure that an iPhone can’t contribute data to your maps? Also, there are some WWW pages that display the locations of known towers. It would be interesting to see tower locations annotated onto your heatmap. If an iPhone app can be used I would be able to start contributing heatmap info.

  18. Duncan says:

    Would it be possible to indicate the ‘freshness’ of the data in view somehow? Something to indicate the time since the last update and estimated time until next refresh?

  19. Duncan Innes says:

    Guys,

    Is there any chance that the relative “freshness” of the data can be indicated on a page? Some indication of how long since the last heatmap was generated for the current map view? Some ETA for the next refresh would also be great. Great site though – doing a great job.

    One last thing I was wondering was if there was some way to indicate the differences between 2G, 3G & 4G on the same map. Could they be plotted with different colour ranges for example? When all three signals are plotted on the same map, it’s not possible to tell which areas are responding to which signal type.

    Thanks in advance

    Duncan

  20. Johan Adler says:

    Regarding updates of data available on the website: I tried this app for a while, but when my data would not show on your site (or in the app) I figured it was not working and I switched to other apps. My current favourite is OpenBMap, simply because my data is visible right away (or at least within 24 hours). Knowing that all my contributions are publicly available is also a strong motivator. Akvelon’s Signal Finder is nice too, the app gives me instant feedback on mapped data. I dislike tha fact that it comes from a commercial company that does not tell me where my contributed data ends up, and their tower positions are completely off in many cases (might be using Google’s data).

    Since I am not the one responsible for making it work I do not hesitate to tell you that the website has to show up to date information, and no user/contributor should have to wait more than 24 hours to see the effect of their work. If you can do that, many more will want to use your app. There are literally dozens of cell signal tracking and mapping apps available, and I might not be the only one testing more than one of them. In order to keep first time users actually using your app, give them/us instant feedback.

    You have a very nice website and beautiful maps, and I will give your app another try. But I think the app should let the user know how to use it, emphasizing the need for “turbo mode” in uncharted areas, and I agree with others suggesting that the default should be that the turbo mode is activated automagically when the phone is charging.

    • Johan Adler says:

      By the way, now I can see that my use of your app a month or two ago actually meant something. The (only) current “heat blob” in this map is right on top of my house. :-)

    • brendan says:

      Johan – just to add. One of the reasons heatmaps don’t get updated so frequently is we are simply having to deal with data on a scale we haven’t experience before. We’ve been working hard on getting on top of this and in the last few weeks have made some real advancements. We’ve just rolled out a full refresh of the heatmaps containing all data collected to date and future refreshes will be much more common.

      We’re also going to be bringing out other delicious features very, very soon :)

  21. Johan Adler says:

    On your first page, somewhere close to the map permalink and the “Under construction” disclaimer (i.e. the ‘seldom updated’ link), please post the dates of the last updates of statistics, tower locations and heat maps.

    On these blog like postings, please show original post date. There is some conflicting information and it would help to know when you wrote something.

    I would suggest that you make a new post explaining data collection strategies, and make sure that there is a post using the current term for “Turbo Mode” (maybe even edit the old one on that subject, updating the title). I spent a long time yesterday looking for the turbo mode setting… ;-)

  22. Digital Diabl0 says:

    I also feel the frustration that others are feeling with the ‘lag’ between the data being sent and it appearing on the map. How about doing like what OSM used to do and have crowd-sourced rendering? If computational power is the limiting factor, why not get others to do the heavy-lifting?

  23. Dan says:

    I have been using the app for about 3 months now. My data still doesn’t appear on your map. When can I expect to see the data?

  24. Matt says:

    Have you thought of letting users voluntarily contribute computational time to generate new averages and heat maps? A little distributed computing could lighten your load substantially, and lead to faster turn-arounds on your data.

  25. Duncan Innes says:

    Is this project dead? There don’t seem to have been any heatmap updates for a couple of months now. You made initial statements that work was underway, but nothing since.

    Can you update us as to what’s going on and when we can expect new heatmaps?

    Thanks

    • Jimbob Yorks says:

      I’m thinking the same. The coverage map has now vanished for me and is now just blank.

      If you’re updating the site then let us know, but equally if you’ve stopped the project let us know too so we can uninstall the app.

      Cheers

    • Bruce says:

      I don’t even see the heat maps anymore. I do see San Francisco (no heatmaps), but if I try to move the location, the map just goes blank. Is anyone seeing heatmaps?
      When will the maps be back up? I have got several updates for the App, and it does seem to be more stable now, but I would like to see the heatmaps.
      Thanks

  26. Robert Smith says:

    Looks like this project is dead – the heatmap is gone.

    Cellmapper is still running strong, and has Android and Blackberry clients.

  27. Tom Bodnar says:

    The heat maps are updated the project is not death. Also several updates of the software were made.

  28. Mike says:

    Can you add a spot somewhere on the map that says “Last Cell Tower Update” and “Last Heat Update”, or something like that? It would be helpful to see when the last updates were to make sure out data is uploaded.

    Thanks!

  29. James Harrogate says:

    An update for your app seemed to be corrupt but despite uninstalling and then reinstalling it it still doesn’t seem to work. It brings up an error message when opening the app and then force closes itself. Do you still offer updates in the UK or are you just concentrating on the US?

    • Sam says:

      Hi James,
      We update the app every wekk or two. The head maps on the website are also update every 1-2 weeks. Both these updates are worldwide – we don’t just do one particular region. If you kit the contact us link at the top of the page you can let us know a bit more about the bug you have and we can hopefully fix it.

  30. Louis says:

    Just wondering when AT&T LTE data will be made available. LTE has been available across most the Chicago area for a few months now. I have been using the app on the Galaxy S II Skyrocket which has LTE and the app seems to recognize all the information correctly. Unfortunately with LTE I never get an approximation of tower location / direction and none of my data I have collected and uploaded in the past week has shown up on the heat maps. There isn’t even an option to view AT&T 4G coverage on the heat map. Any estimate when this data will be available?

  31. Kellogs says:

    How long does it take to display new tower location. Do you get the list from operators or do you create the list using triangulation? Kind regards from France.

    • Jon A says:

      I would assume triangulation, and it’s not overly accurate. Here in Australia the sites are up to 20KM out, and there’s a list from the federal regulator of all cell sites.

  32. Radioactive Man says:

    My data from over a month ago is still not appearing (Telstra and Vodafone in Canberra, Australia). I have been making a very special effort to contribute over this period. Will it ever appear? The map does not seem to change at all over many months.

  33. Jon A says:

    Hey guys! I know that your data is all submitted and randomly, but many of the tower locations are completely off in these parts. I do, however, have a site with all the locations listed. Are you guys interested?

  34. Hans says:

    I walked a not covered area and forced the upload one month ago, but there is still no coverage in that area. Could you be so kind and include the “last heatmap update” in the map somewhere?

    Thanks

  35. Andrew Steer says:

    The cell tower locations are out by up to half a mile or more in my part of the world (Cambridge, England).
    I see people have commented on this before. Can you explain how you figure out the tower locations, and maybe I can put some brain cycles to thinking if there’s a better way.

  36. ECPS92 says:

    I’ve been adding Sprint info for months [Dec-2011 to present] and still many towers are not uploaded for the Mass/NH/RI area of the US :(

    If you have given up on the project let us know and we can move onto a project site that is actively supported.

  37. Wellington says:

    I’m impressed with RootMetrics’ Android app, the network information that I collected was included on the map in a matter of minutes!

  38. anderson110 says:

    Alright, the goal of daily updates was set over 2 years ago. Are you there yet?

    If not, why not? Monthly updates is a bit of a joke.

    • paemt610_7 says:

      Promise of daily updates? Indeed above it was set as a goal in 2011. I have been using the app since at least April 2014, this is October 2014 and I have yet to see any recent data added to the heat maps. I have traveled the same path to work 4 days a week since April with no data added to the map for that path. Visiting family once a week in another path also has yet to appear with data, and both of these paths were blank when I started and still remain to be blank. I have used other signal mapping applications that have the ability to update the map within 24 hours of me collecting data, some within a few hours literally less than 3 hours of collecting the data. I would imagine that OpenSignal needs to research a more efficient way of utilizing the collected data into something useful. What is the point in collecting data that takes virtually years to process and make useful? Why are other apps that do the same service able to reduce the same worldwide point processing into hours instead of 6+ months?

  39. swordfishBob says:

    How about some way of recognising data that’s substantially different to existing averages, or especially where there’s no existing record, and flagging that for more rapid calculation?

Leave a Reply