20 October 2010 at 09:11 #495
Hello – we are working on to support the recent Hungarian enviromental crisis. Charities has already prepared an excel sheet, with adresses. We would like to put them on a map – is there any good practice/short way to do this?
Csaba20 October 2010 at 12:18 #3053
Check how the CSV file is prepared for uploading huge data to Ushahidi.
#,INCIDENT TITLE,INCIDENT DATE,LOCATION,DESCRIPTION,CATEGORY,APPROVED,VERIFIED
“1″,”Suspected death in Nairobi”,”2009-05-15 01:06:00″,”Nairobi”,”Three cases have been confirmed in C. del Uruguay”,”DEATHS, CIVILIANS, “,YES,YES
“2″,”Looting”,”2009-03-18 10:10:00″,”Accra”,”Looting happening everywhere”,”RIOTS, DEATHS, PROPERTY LOSS, “,YES,NO21 October 2010 at 07:26 #3054
thnaks for the informtaion!
I think, this will be useful. We would like to use ushahidi to map the existing donations, where they are, what are these material donations – and to manage the logistics to pick them up.
Csaba29 October 2010 at 03:06 #3055
I am also trying to do a batch upload using the “upload reports” feature. I can get everything to upload except for the location on the map. I need to be able to plot a path of a tornado and I would like to try using Ushahidi as a tool. Please advise.
Lee11 November 2010 at 12:34 #3056
I succeeded in importing 2000 pre-geocoded stories via CSV upload.
It was unquestionably an undocumented & tedious task with mostly manual error correcting. The bulk upload system doesn’t allow UNICODE and my data had special characters. So it would crap-out on the first line and I would have to go in and find the weird character and re-upload it again, over and over. If it could have highlighted all the errors, that would be an slight improvement.
The trick is to add two more columns not specified in the directions: LATITUDE and LONGITUDE. They must be in caps and filled in, and VERIFIED must be YES for them to appear, otherwise you have to manually approve 2000 reports. (please shoot me!)
So all said, ushahidi was pretty much a terrible experience for managing data in bulk, since you couldn’t click a “select all” and hit verified =YES for 2000 reports. And it doesn’t batch-geocode reports based on fuzzy semantic logic like batch Geo does. I would have found it easier to hack into the MySql and run a query to do manage bulk processing, but I’m rare in being able to use mysql/php somewhat.
BTW – someone on the ushahidi tech team DID point me to a bulk geocoding database you could use for your data before you upload it: geonames.org. There’s also “batch geo” – but didn’t work in Kenya.
I ended up manually identifying 30 or so locations in google maps, and writing a python program to match these locations to reports, and inserting the LOCATION/LAT/LONG into the CSV. Yes -this is another way that the management interface requires a lot of expertise. I’m rather surprised to be hearing the “u” word at half the conferences in Washington these days when it is still a product in development from my perspective as a data manager. But then again none of those who talk about it in conferences have ever used it to map something. They simply love the slick public appearance of the data. (They’re more like fans, not players)
(I’m a little pissed that I have to redo all of this and the latest version won’t install because it requires a php upgrade that bluehost doesn’t intend to make)11 November 2010 at 16:10 #3057
hm, this is not so good news! I appreciate your kindess sharing your experiance – even it is not so uplifting.
However, do you think it could be a workable solution, to connect with a simple openoffice database through an sql connector like http://extensions.services.openoffice.org/project/mysql_connector directly to the datbase?
We also need unicode characters very much!
The forum ‘Troubleshooting’ is closed to new topics and replies.