Brandwatch data is currently only accessible to Brandwatch’s customers; those with a login to our Analytics platform. This powerful data analysis tool is not free for public use, so the data on which it is built is not easily shareable. But we want to change this.

We’re not in a position to open up the app for free. Not yet anyway (we’re working on it). But there are still ways we can share some of what our queries return.

Below is some of the data we have extracted for various React, Data Labs and other projects (using the Brandwatch Analytics API). The projects are presented “as is”; with no edits, cleaning or documentation.

The data is for you to explore and do with what you will. All we ask is a credit if you use it for anything groovy. Thanks.


Brighton Digital Festival (Sept 2015)

bdf2015-opening-BDF15photogruff_PGF3926

Now in its fifth year, the Brighton Digital Festival is a month long celebration of the city’s digital culture. Brandwatch sponsors the festival, takes a role in the governing consortium and also, on occasion, creates an event too (see “For The Win(dow)” below).

We also track the social conversation around the festival too of course. The XMLs below contain a load of data on the 2014 and 2015 festivals.

Volume of mentions

1 – 27 Sept 2015: http://labs.brandwatch.com/mattp/DATA/BDF/volume/FixedPeriod_2015byday_BDF.xml

The equivalent for 2014 (1 – 28 Sept): http://labs.brandwatch.com/mattp/DATA/BDF/volume/FixedPeriod_2014byday_BDF.xml

The longer view, July to October
2015: http://labs.brandwatch.com/mattp/DATA/BDF/volume/FixedPeriod_2015threemonth_BDF.xml
2014: http://labs.brandwatch.com/mattp/DATA/BDF/volume/FixedPeriod_2014threemonth_BDF.xml

Topics of conversation

2015
September – http://labs.brandwatch.com/mattp/DATA/BDF/topics/FixedPeriod_2015byday.xml
Jul-Oct – http://labs.brandwatch.com/mattp/DATA/BDF/topics/FixedPeriod_2015threemonth.xml
Week by week – 31st-6th / 7th-13th / 14th-20th / 21st-27th

2014
September – http://labs.brandwatch.com/mattp/DATA/BDF/topics/FixedPeriod_2014byday.xml
Jul-Oct – http://labs.brandwatch.com/mattp/DATA/BDF/topics/FixedPeriod_2014threemonth.xml
Week by week – 1st-7th / 8th-14th / 15th-21st / 22nd-28th

Most impactful authors

2015
September – http://labs.brandwatch.com/mattp/DATA/BDF/authors/FixedPeriod_2015byday.xml
Jul-Oct – http://labs.brandwatch.com/mattp/DATA/BDF/authors/FixedPeriod_2015threemonth.xml
Week by week – 31st-6th / 7th-13th / 14th-20th / 21st-27th

2014
September – http://labs.brandwatch.com/mattp/DATA/BDF/authors/FixedPeriod_2014byday.xml
Jul-Oct – http://labs.brandwatch.com/mattp/DATA/BDF/authors/FixedPeriod_2014threemonth.xml
Week by week – 1st-7th / 8th-14th / 15th-21st / 22nd-28th

Demographics

2015
September – http://labs.brandwatch.com/mattp/DATA/BDF/demographics/FixedPeriod_2015byday.xml
Jul-Oct – http://labs.brandwatch.com/mattp/DATA/BDF/demographics/FixedPeriod_2015threemonth.xml
Week by week – 31st-6th / 7th-13th / 14th-20th / 21st-27th

2014
September – http://labs.brandwatch.com/mattp/DATA/BDF/demographics/FixedPeriod_2014byday.xml
Jul-Oct – http://labs.brandwatch.com/mattp/DATA/BDF/demographics/FixedPeriod_2014threemonth.xml
Week by week – 1st-7th / 8th-14th / 15th-21st / 22nd-28th


The Labour Leadership Election (May – Sept 2015)

We tracked mentions of the four leading candidates – Andy Burnham, Yvette Cooper, Jeremy Corbyn and Liz Kendall – from May to September, and visualised their relative volume and sentiment here.

The raw volume data behind this is encapsulated by these four XMLs:

https://labs.brandwatch.com/mattp/DATA/labourLeadership15/volume/FixedPeriod_maySept_kendall.xml

https://labs.brandwatch.com/mattp/DATA/labourLeadership15/volume/FixedPeriod_maySept_corbyn.xml

https://labs.brandwatch.com/mattp/DATA/labourLeadership15/volume/FixedPeriod_maySept_cooper.xml

https://labs.brandwatch.com/mattp/DATA/labourLeadership15/volume/FixedPeriod_maySept_burnham.xml

We didn’t use it for our visualisation, but we also gathered data on the top topics being discussed across social, broken down by both sentiment and candidate. That is here:

https://labs.brandwatch.com/mattp/DATA/labourLeadership15/topics/Topics_FixedPeriod_maySept.xml


Glastonbury (June 2015)

glasto2

During the glastonbury weekend we tracked mentions matching:

glastonbury OR glasto OR hashtags:glasto OR hashtags:glasto2015 OR hashtags:glastonbury OR at_mentions:glastofest OR "worthy farm" OR worthyfarm OR eavis

I then extracted the topics of conversation from the mentions returned, and graphed them here. The topics were tracked day by day and hour by hour, and can be accessed using this format:

https://labs.brandwatch.com/mattp/DATA/glasto15/topics/Topics_YYYY-MM-DD.xml

… for days, and …

https://labs.brandwatch.com/mattp/DATA/glasto15/topics/Topics_YYYY-MM-DD_HH.xml

… for hours. Hours are 00 to 23, and are in GMT (one hour earlier than the stage times). Data includes a sentiment breakdown of the topics too, which wasn’t used in the viz.

For example, if you want top topics while Kanye was onstage, you’ll be wanting 11pm (10pm GMT) on Saturday 27th. Which is…

https://labs.brandwatch.com/mattp/DATA/glasto15/topics/Topics_2015-06-27_22.xml


UK General Election – The Debates (April 2015)

Love them or hate them, the leaders debates are once again key focal points for the 2015 election campaign. We’re tracking live volume and sentiment data for each of the leaders in each of the debates.

NB 1 – The data will be available via the links below AFTER the debates. If you want live data, you should probably talk to our sales team.

NB 2 – all times are GMT.

26th March – Battle For Number 10

https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate1_cameron.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate1_miliband.xml

2nd April – Leaders Debate

https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate2_bennett.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate2_cameron.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate2_clegg.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate2_farage.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate2_miliband.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate2_sturgeon.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate2_wood.xml

16th April – Challengers Debate

https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate3_bennett.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate3_farage.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate3_miliband.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate3_sturgeon.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate3_wood.xml

30th April – Question Time Special

https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate4_cameron.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate4_clegg.xml
https://labs.brandwatch.com/election15/opendata/FixedPeriod_debate4_miliband.xml


UK General Election (January-May 2015)

Starting with the EU Elections in May 2014, I’ve been tracking Twitter mentions of the main UK political parties with view to seeing if an election result could be predicted. With occasional success. I’ve written up my experiments here.

The data is continuing to gather for the 2015 General Election, and data from 1st Jan is now available to share.

How to find the data

To retrieve data you just need to know the filename. The format of my naming is as follows:

https://labs.brandwatch.com/mattp/DATA/UKelection/volume/[date]_[party].xml

[date] is in the format YYYY-MM-DD (e.g. “2015-01-12” = 12th January). Earliest available date is “2015-01-01”.

[party] is either “GREEN”, “LABOUR”, “LIBDEM”, “TORY”, “UKIP” or “SNP”.

So, for example, all UKIP data for 4th January 2015 is at

https://labs.brandwatch.com/mattp/DATA/UKelection/volume/2015-01-04_UKIP.xml


Brighton Digital Festival, For The Win[dow] (Sept 2014)

Throughout the month of September 2014, people were tweeting pixels at International House as part of the 2014 Brighton Digital Festival. All valid tweets, and the designs they contained, were stored for use on the gallery page, and so can be shared here. For more info about the project, see here.

The data is all in one big text file, at

https://labs.brandwatch.com/FTWindow/twitterdata/frames.txt

Each line is in the format:

[unix timestamp] [twitter id] [target window] [data]

e.g. 1410463689 davidlees44 01 FFFFFFCF868686CD034F4F07B7B7B733 FFFFFFCF858585CD034F4F07B7B7B733 FFFFFFCF838383CB034F4F07B7B7B733

.. was a tweet from @davidless44, sent at Thu, 11 Sep 2014 19:28:09 GMT, intended for window pane number 1.

The graphical data, in hex form, needs to be converted to binary. Each image is represented by a 32 character string, 16 two-digit hex numbers, which can be converted into a 128 character binary string. This can then be split into chunks of 8, to make the 16 rows of the 8×16 frame. Below is a Processing.js function that does just that, as an example:

int PIXEL_SIZE = 6;

void decodeStringToCanvas(String string, int ex, int why) { 
     if (string.length() == 32) {
	  int y = 0;
	  for (int i = 0; i < string.length()-1; i += 2) {
    	            int x = 0;

		// read two hex chars
		String hexstr = "" + string.charAt(i) + string.charAt(i+1);
		int c = unhex(hexstr);

		// convert that number into 8 char binary string
		String eightBits = binary(byte(c), 8);

		// loop through it
		for (int j = 0; j < eightBits.length(); j++) {
		  int val = int(eightBits.charAt(j));
		  if (val == 1) { 
		  	// draw a circle for every “1” 
			ellipse(ex + (x * PIXEL_SIZE), why + (y * PIXEL_SIZE), PIXEL_SIZE, PIXEL_SIZE);
		  }
		  x++; 
		}
		y++;
	  } 
     }
}

[EDIT: Python implementation, by @williamsmj_ here: https://github.com/williamsmj/ftwindow.py/blob/master/ftwindow.py]


Scottish Referendum (August/Sept 2014)

scotland

Tracking YES/NO conversations in the run up to the 2014 Scottish Independence Referendum.

Following a surprisingly successful attempt to predict the 2014 European Elections, I'm currently tracking data on the upcoming Scottish Referendum to see if the trick is repeatable. An explanation of the logic behind the query and a few graphical representations of the data are here.

These are the raw data files:

http://zenbullets.com/BW/EUelections2014/SI_CategoriesByDay.json
http://zenbullets.com/BW/EUelections2014/SI_CategoriesByDay_justScotland.json
http://zenbullets.com/BW/EUelections2014/SI_CategoriesByHour.json
http://zenbullets.com/BW/EUelections2014/SI_CategoriesByHour_justScotland.json

These files will continue to be auto-updated on the hour until the close of polling on Thursday 18th September 2014.


Commonwealth Games (July/August 2014)

Data accumulated for an uncompleted viz. Structure of the data is very similar to the World Cup project (below), each file containing total tweets for a certain time period. But breakdown is on sport rather than country.

How to find the data

To retrieve data you just need to know the filename. The format of my naming is as follows:

https://labs.brandwatch.com/commgames2014_data/[TYPE]_[datetime][_sport].xml

The [TYPE]s available for any given period are:

"TOPICS" - top topics

The [datetime] is written in the following format: YYYY-MM-DD_HH. So, for example, "2014-06-03" is the 3rd of June, "2014-06-03-15" is the 3rd of June 15:00-16:00 GMT.

The [_sport] (which is optional) must one from the following list:

"aquatics", "athletics", "badminton", "boxing", "cycling", "gymnastics", "hockey", "judo", "lawnbowls", "netball", "rugbysevens", "shooting", "squash", "tabletennis", "triathlon", "weightlifting", "wrestling"

So, for example, the filename to retrieve top topics around the subject of badminton between 13:00 and 14:00 GMT on July 22nd, would be:

https://labs.brandwatch.com/commgames2014_data/TOPICS_2014-07-22_13_badminton.xml

Gather started on 10/11th July, and is currently still running. First full day is 2014-07-10, first full hour is 014-07-11_10.

For example:

Top topics, first full day - https://labs.brandwatch.com/commgames2014_data/TOPICS_2014-07-10.xml

Latest / Live Data

While the gather is running, you can also access data for the last hour, last day, last week, or since the start date. These files are at a separate location. The format is similar:

https://labs.brandwatch.com/commgames2014_latest/[TYPE]_[timeperiod][_sport].xml

[TYPE]s are as above: "TOPICS"

[timeperiod] is either "lastHour", "lastDay", "lastWeek", or "total".

The [_sport] (which is optional) is, again, from the same list as above.


Football World Cup (June/July 2014)

wc_data

This is the data accumulated while our World Cup 2014 Data Dashboard was live.

Each XML file contains total tweets for a certain time period. The totals are twitter mentions of world cup topics, with a further breakdown down on mentions of each country in the competition. The data was gathered using a 1/10th sample, which was then multiplied up. This is why all values are multiples of 10. If these totals disagree with Twitter's reported volumes it is because Brandwatch are considerably more ruthless in their criteria for relevance (removing spam, duplicates, etc).

How to find the data

To retrieve data you just need to know the filename. The format of my naming is as follows:

https://labs.brandwatch.com/worldcup2014_data/[TYPE]_[datetime][_country].xml

The [TYPE]s available for any given period are:

"TOPICS" - top topics

"MONITOREDTOPICS" - top topics from a predefined list

The [datetime] is written in the following format: YYYY-MM-DD_HH. So, for example, "2014-06-03" is the 3rd of June, "2014-06-03_15" is the 3rd of June 15:00-16:00 GMT.

The [_country] (which is optional) must one from the following list. Note the "_"s and lowercase:

"australia", "iran", "japan", "south_korea", "algeria", "cameroon", "ghana", "ivory_coast", "nigeria", "costa_rica", "honduras", "mexico", "united_states", "argentina", "brazil", "chile", "colombia", "ecuador", "uruguay", "belgium", "bosnia_and_herzegovina", "croatia", "england", "france", "germany", "greece", "italy", "onetherlands", "portugal", "russia", "spain", “switzerland"

So, for example, the filename to retrieve top topics around the subject of germany between 16:00 and 17:00 GMT on June 28th, would be:

https://labs.brandwatch.com/worldcup2014_data/TOPICS_2014-06-28_16_germany.xml

Date range is approx 3rd June - 14th July. First full day is 2014-06-03, first full hour is 2014-06-04_15.

For example:

Top topics, first full day - https://labs.brandwatch.com/worldcup2014_data/TOPICS_2014-06-03.xml

Last full day is 2014-07-13. Last full hour is 2014-07-14_14.