I have been helping clients design, build, and deploy data platforms for many years. Initially on premise, but now in the cloud. I think your choice of technology is less important than your chosen data architecture. I would focus on defining this data architecture first (i.e. to satisfy your requirements) and then looking at how the Azure, AWS, or GCP clouds could best satisfy your requirement. I would consider things like data ingestion, orchestration, modelling, and exploitation. The biggest challenges won't be technology, but thinkings like 1) how do you model the data to answer your questions and 2) how do you relate/join the various data sources. I hope this helps; feel free to pop in a call of you would like to delve into more detail or if you have any questions.