Ghana: Child Slavery Campaign - Why does DG use older images?


I’m not having any problem so far with disappearing tags since I’m obviously in the middle of the lake. The one problem I have is that the image is 5 years old. I can understand the need for a baseline, but … this seems excessive.


Nodder Introductions
Ghana: Child Slavery Campaign - Disappearing tags

@HappyMapper I think that it is random… usually it is January 2014 on the bottom and I don’t care which campaign it is the date is almost always the same. So I am thinking that maybe it is random… or perhaps it is trying to do a comparative study as to increase of population with a bigger housing count, or more boats… it is speculation but it does seem logical. :coffee:


Hi @Harold

Oftentimes, it seems DG’s clients want a retrospective count and a recent count, like the comparative data that @Pastor_Ruth mentioned. i suspect some campaigns can pull up an area from the billions of images in their archives without concern about the freshness-use by date?



I’m relatively certain that those dates are not random. What you see is what you get. I agree with cageycat’s interpretation.



For this (and most campaigns), we use several different images in order to create an image mosaic covering the entire area. In the Ghana: Child slavery campaign there are a few dozen individual images. In all campaigns where image date is important, we use the most-recent cloud-free imagery available. Urban areas tend to have a higher level of demand, so much of the satellites’ time and image holding capacity is spent on those higher-demand areas.

The imagery used in the Ghana: Child slavery campaign are the most-recent cloud-free images we have in order to achieve full-coverage of the AOI. There are some areas where there is overlap between a newer and older image (for example if there is a gap between fresher images in the Area of Interest (AOI) we might need to go back a few years in order to get an image. A newer image might overlap with part of that older image). Date will be taken into account in the research. However, the majority of imagery in this campaign is much more recent.


Unfortunately, every time I log in and try to work this campaign all of the images offered date to December 2010. I have yet to see anything more recent (and believe me I’ve tried.)



Hi @Harold

When I think about date issues and concerns, I try to put it into a different context. For example, we all know that New York City has a high percentage of runaway kids and homeless teens and adults. That trend probably doesn’t change much in a 5 year period. Let’s say Tn or a client-contract identified that most homeless people tend to camp or sleep near a certain kind (shape) of warehouse (perhaps ones with pitched roofs instead of flat roofs). Let’s say Tn also wanted to count the tallest buildings with a triangle shape on top. And just to add some strangeness, tag every street-intersection and any campfires/barrel fires the homeless make on a cold night. (just making up these details.)

Now, none of those would change much in 5 years. We’d not be able to count numbers of people–that would need done on the ground. But a building - warehouse count would be useful. Might not see many barrel-fires, especially if there are clouds. Street intersections might change, but not much.

Ghana’s structures and infrastructures wouldn’t change that much either. Cultures, anywhere, tend to stay fairly stable over time unless during some in-country strife (ex. droughts) or war especially with mass migrations. So perhaps Ghana villages add 1 or 2 buildings per a 2-5 yr period. Long boats fail/break and new ones built, but the total stays relatively at the same level-- unless someone like Trump donates a billion dollars for people to buy fleets of long boats. Same with fish cages that wear out, break, or sink. But as long as the ocean / water and shoreline remains pretty much the same, villagers will keep living and working in the same areas decade after decade, generation after generation-- unless the youngest generations find ways to get out, move.

Social scientists learn to extrapolate and predict changes in “number” facts, like changes in population, or how many people live in 1 hut or house (correlates with ground census), etc. etc.

So while we are working on 5 yr old images, the data is still valid as a snapshot in time. For example, I’m surprised I’m not seeing fleets of boats, and big boats, absent at least in the images I’ve done. Lots of cages, but few boats. So I wonder what conclusions social scientists will draw from these counts and … how the puzzle fits together??

There’s a ton I don’t understand about studies, especial quantitative (numbers). But I know that much can be learned even from a small amount of data, and even if 5 yrs old.

On another front, it’s more cost effective for a (probably) underfunded org to request Tn to use historical images. Many orgs could never afford to contract a new tasking mission, at least that’s my view/understanding. Plus, if every new contract meant tasking one of the satellites, Tn couldn’t sustain the financial burden–again, just my view/ opinion/ assumption. Plus, poor WV3 would get tuckered out zooming here and there to constantly take new pictures. It isn’t like he can stop at the big Satellite Repair Shop in the sky for a tune-up if his imager develops hiccups or heartburn. So, constantly tasking satellites for many new missions is likely not feasible! They are taking pics every day, I think.

As I said in other posts, I trust that the people at DigitalGlobe know their stuff and take all the particulars into account when choosing the images, coloring, new vs historical, etc. IF the data would be useless to social scientists, Tn would know that in the planning stage. They aren’t doing this work to produce useless results.

However, that said, our review of images might not find what was expected-- and in science, even “negative findings” are a very useful “result”. I think that’s what makes this kind of research exciting-- never know how a research question will be answered… except that there will be some answers. For example, Cocos Island… we didn’t find what was expected, but the data was still facts that go into making the next research question… and the next…and the next.

Don’t worry if images seem “old”. Some research doesn’t need brand new pictures, but just studying what they already have. Trust that DG knows what they’re doing.


Please tell me why the satellite images for Ghana are dated January 2010 – over 5 years ago.


Hi vicmep, check the explanation Happymapper gave some days ago :grinning: