This posting focuses on the quantity of Fotopedia photos, while a previous posting focuses on the quality of Fotopedia photos. It is based on daily “measurements” starting on September 17th 2010 and currently covers almost 6 weeks.
The main question I am trying to answer is whether Fotopedia is on-track to becoming a major site on the Internet. For that I want to know if the user community and the rate of photo submissions is growing, stable or even decreasing.
I simply recorded the statistics displayed on the Fotopedia website. I did this on a daily basis at roughly the same time of day: evening in the Western European time zone when site activity seems to quiet down a bit.
Number of photos
The amount of photos in the database is currently about 590k (thousands) and grows at about 2% per month. Continued linear growth at this rate would give about 25% growth in the coming year (840k photos by 1-Oct-11). This may seem a lot, but is IMO not necessarily enough to become a lasting phenomena: such phenomena require exponential growth until they get to the “household word” level.
This is consistent with Fotopedia’s ranking in Alexa.com of in the order of 35,000 – meaning that the site belongs to the top fifty thousand sites in the world (based on estimated traffic during a 3 month period). Wikipedia and Google for example are in the top ten. For comparison, a brand that only indirectly depends on the Internet like Skoda-Auto.com ranks 46,267 and a largely Internet-based service like GettyImages.com rates 1,933.
The dotted line in Fig. 1 represents the 4.4% of the photos that are part of the World Heritage Sites project. These Heritage photos are growing at only 1% per month – likely because the collection is closer to saturation: most Heritage articles are already covered and there has been a campaign to get enough photos for the Heritage application in time for the first release of the App on iOS (a 2nd version has just been released).
Number of articles
The database currently covers 41k articles from the English-language version of Wikipedia.
On average each article has about 14.3 photos. This photo/article ratio seems stable, but is an average across some articles with hundreds of photos, articles with dozens of photos and articles with only a few photos. Articles with no photos are typically not counted as they don’t occur in the database (although there are temporary exceptions).
The amount of articles in Heritage has varied between 3771 and 3774 during the monitoring period – and is thus extremely stable. The variations only reflect refinements because the UNESCO list used as a basis is seldom extended. The UNESCO list contains 911 sites or groups of sites (e.g. the Historic Center of Cordoba corresponds to multiple Wikipedia articles).
For Heritage, the photo/article ratio is about 6.9 and is slowly growing: the number of photos grows while the amount of articles stays constant. This 6.9 ratio is about half of the average of 14.3, possibly because many heritage sites are in exotic locations and because the Heritage project is carefully curated.
Another way to look at these numbers: 10% of the Fotopedia articles form the Heritage project, but only 4.4% of the Fotopedia photos form the Heritage project.
Is a total of say 40k articles a lot? Well, the English version of Wikipedia is expected to reach 3.5 million articles by the end of 2011 (and all Wikipedia languages together may give 7M unique topics). So Fotopedia covers roughly 1% of the scope of en.Wikipedia.org. The coverage is probably much higher for Wikipedia articles on well-known places (Palace of Versailles), cuddly animal species (koalas, porcupines), celebrities and fandom (Elvis, Arsenal FC, BMW), some arts (architecture) and common objects (pen, mailbox). And is lower for advanced topics like science (Cobalt, string theory), mathematics (differential equations), technology (strain gauge – my photos!), history (Treaty of Versailles), society topics (bar is still missing), and pretty much anything that is abstract (octave).
Photos added per day
Figure 3 shows the change in the total number of photos on a daily basis. On average about 500 photos are added per day. It is thus the same data as Figure 1, but with the small daily changes magnified. On some occasions there was a decrease – but I was told this was because of semi-automatic removal of low-rated images by Fotopedia staff.
The peak of +3000 photos in one day is notable, but I couldn’t find out what caused it.
Based on this limited amount of data, there is no clear evidence to show that the rate of growth is either picking up or slowing down. So for now, I assume a more or less linear growth at roughly 500 photos/day.
Articles added per day
The database grows at roughly 40 articles/day. Note that, 500 photos and 40 articles per day is close to the 14.3 photos/article ratio. So the current photos/article ratio is not shifting to higher or lower ratios.
The current size of Fotopedia (1% of Wikipedia topics, average of 14.3 photos/topic) is in itself OK – given that Wikipedia is full of specialized topics. The rate of increase (500 photos/day) is also not a problem in itself.
But the available data suggests that at present a more or less stable community of active people is providing data at a constant rate. This constant growth rate is probably not enough to boost Fotopedia to become a major Web 2.0 service.
Hopefully the future Fotopedia iPad apps and other planned site improvements can help attract more attention for Fotopedia.
This conclusion appears to be consistent with Alexa’s statistics (see Fig. 5) that show that the rapid growth characteristic of a new Internet service pretty much stopped in early 2010.
Note that although Alexa numbers can easily be challenged on various grounds, the Alexa ranking is accepted as a quick and dirty way to assess the importance of a web site. In this particular case, there is an extra complication: access to Fotopedia photos via the Heritage app is simply not counted by Alexa because Alexa basically monitors browser activity (using an optional toolbar).
Note that the time axis of the Alexa data is much larger than the other graphs.