From FreeMalaysiaToday, July 14 2011 (http://www.freemalaysiatoday.com/2011/07/14/bersih-rally-maths-puts-crowd-at-4500-to-50000/)
A research firm using tweets, photos has charted a way to measure just how many people turned up at the ‘Walk for Democracy’ rally.
PETALING JAYA: Crowd estimates for the July 9 Bersih 2.0 rally are almost as contentious as the rally itself. Various figures have been bandied about. Just about everybody has their own estimation – the organisers, police and the media.
Now, one man has come out and said that with his mathematical calculations, the figure is around 45,000 and 50,000. He is Politweet founder, Ahmed Kamal Nava.
Politweet.org, a research organisation that studies Twitter interactions between Malaysian citizens and politicians, had mapped out the number of people who attended the rally using polygons.
But before proceeding to draw the polygons, a timeline of events needed to be established.
Ahmed Kamal told FMT that his primary source to determine the timeline of events were tweets and also pictures posted on Twitter as these pictures, unlike the ones posted on Facebook or other websites, had the integral time stamped on them.
“Using the images from the photo gallery as a reference, I started to draw the polygons covering the area on a map. I used some photos found online, if they were of higher quality, and matched them up with my established timeline,” he said. He established that at about 1pm to 2.30pm the crowd was at its largest.
“The crowd estimation is based on this peak period for different areas within the same time frame. This is to avoid double-counting because crowds were moving, growing and shrinking between 1pm and 4pm,” he said.
“Using the polygons, I was able to use a custom tool to estimate the covered area in square feet. With the factors stated on the website, it was possible to gauge how many people were gathered.”
“Polygons were then generated based on the photos taken within that time frame in different parts of the city,” he said, explaining the methodology used on Politweet’s website.
“If the crowd was standing close together, the factor was one person for every 4 sq ft (one person/4 sq ft). If the crowd was moving about, or a mix of people standing close and far, then it was one person for every 5 sq ft to 6 sq.ft. The factors used are stated for each map.”
He had focused on three areas of the city during the peak period – Puduraya, KLCC and Sungei Wang Plaza. The other “hot spots” such as Pasar Seni were either counted as a subset of these three areas or fell under the “others” category.
“Crowds that were close to Puduraya (such as at the Agro Bank at 12.30pm) are assumed to have joined the Puduraya crowd,” he said.
“Smaller groups were recorded at other areas such as Jalan Raja Laut, Pasar Seni, and Jalan Maharajalela and are all assumed to be within the ‘other’ group of 2,000. This also includes people who stayed behind in Petaling Street when the majority moved to Puduraya at 1pm,” he added.