Local aggregations to improve data access on singe pod, a mobility application perspective

pbonte commented 2 years ago

Pitch

Public transport providers, such as de Lijn, are interested in an overview of the most used bus stops. As each user can log their location traces, these apps can issue time-consuming queries to the user pods in order to aggregate the location traces and count the users that hopped on or of at each bus stop. However, evaluating these queries is time-consuming as the whole location trace needs to be investigated. Therefore, location aggregation can be precomputed that prepare these location summaries.

Desired solution

A local aggregator that can precompute the location summaries of each user, based on the issues queries of the mobility applications. This requires to:

Investigate the application queries in order to determine how the local aggregation needs to be computed
update the aggregations/location summaries when new location traces have been produced
efficiently store these aggregations locally and redirect the queries to not use the whole dataset

Acceptance criteria

A demo that showcases this solution would need to be able to:

show the speed up loading time of using the aggregator compared to no aggregator
show that the location summaries get updates when contacts post a new message
aggregations are stored locally on the data pod

pheyvaer commented 2 years ago

For a demo to be implemented, do you have data that can be used for this?

pheyvaer commented 1 year ago

@pbonte Did you have the chance to look into making the necessary changes?

SolidLabResearch / Challenges