App Annie - API
App Annie Apple Gives Access to three data product in Knoema.
App Annie - Usage Product Level
App Annie - Store Product Level
App Annie - Store Rank Product Level
Data Dictionary Link: https://helpcenter.appannie.com/community/s/article/App-Annie-Bulk-Data-API-Documentation
Creds: Ask May.
Product level is a reference to “product_id” used as the main key. Metric bundles link to dimensions, countries, categories. All joins have been done in OASIS.
@Former user (Deleted) Hi Hari, as you and Graham suggested, I have summarized everything I have and listed them here for whoever is going to help take this over.
@Emily Somma (Unlicensed) @Graham Tolson (Deactivated) Please feel free to add more details if I’m missing any.
Some collections from previous work:
Please apply for AN AppAnnie account from the vendor for graham/hari. Before that you could ask me for my account and I’d love to share. But I still strongly recommend reaching out to AppAnnie for your own accounts.
Current datasets built:
-AppAnnie Usage Product Level Unified Product - App View
https://knoema.com/APPANNUNPAPP/appannie-usage-product-level-unified-product-app-view
2019.01.01 - now
-AppAnnie Store Product Level Unified Product - App View
https://knoema.com/APPANNSPLAPP/appannie-store-product-level-unified-product-app-view
2019.11.01 - now
Did not load more history because of painful joins due to large data, as well as Dataops failures (please check this link for reference) which have just been fixed on Apr 29th. But I believe bringing in history will be better after we restructure the datasets.
Client wants 3 years history in total.
Requests from client (Specifically for Usage Product dataset and Store Product dataset)
(0)The ideal dataset structure is expected to be able to reproduce what AppAnnie has on their UI under <Unified Analysis>. Here is an example. The current structure is able to fulfill most of the functions, only need to lose 2 dimensions and add 1 metric (listed in (1) and (2) below).
(1)Lose Parent category and category dimensions
(2)Bring in Unified Analysis (top level) aggregation numbers
(3)Ingest only US region companies/apps
(4)Indicators that client cares about:
i.Downloads - store_product
ii.DAU - usage_product
iii.WAU (I don’t think this is available from AppAnnie, if client really needs this specifically, please check with vendor whether they have it)
iv.MAU(I don’t think this is available from AppAnnie, if client really needs this specifically, please check with vendor whether they have it)
v.Time per user - usage_product
(5)Current we only ingest delta files because spark cannot handle that many joins on such big history files. Please decide whether we want to ingest all-history instead of delta file (changing append mode into overwrite) going forward, since we’re trimming the data size a lot now.
(6)- Oasis links: http://oasis.knoema.com/data_service/etl/1db0c2bd-89ae-4799-a6d8-79ba46f8ca91
http://oasis.knoema.com/data_service/etl/0a577aaa-a5cb-444d-8c35-60519d0b352e
- Transform examples with delta data: http://emr.knoema.com:8890/#/notebook/2G5NWA7EU (If have no access please ask sysops for it)
- File locations on S3: s3://knoema-prod-data-lake/active-historical/prod/data_provider=appannie/
(p.s. <active-historical> is an S3 location where “60-day removal rule” does not apply, so all the files will not be removed. If have no access please ask sysops for it)
(Not important at the moment but need to circle back after all else are done) After the re-ingest for the datasets above are done, please check with Emily whether client needs the rank dataset(also needs modification) https://knoema.com/APPANNSRPL/appannie-store-rank-product-level For now, it can wait for finishing everything else first.