TensorFLow : google’s machine learning api can be super powerful.
Just make json REST calls to the end-point and get results based on google’s machine learning lib.
Uses cases:
1. identify an image (image classification)
2. parse speech into text
3. translate languages
4. use for predictive analytics to run wide and deep learning on datasets
Google Cloud Platform : It is Google’s AWS cloud.
Key takeaways: They only charge by the minute and not by the hour like amazon.
Extremely cheap pay as you use model on an extremely fast network; Access to Google’s pipes and servers.
A ton of of useful Cloud tools: SaaS, PaaS, BaaS etc..
Compute Engine : AWS EC2 boxes except more scaleable.
App Engine : AWS Beanstalk
Cloud Storage : AWS S3
data tools : the ones I’m interested in.
Google’s Big Table : Similar to amazon DynoDB, a nosql data store. The basic store engine of gmail,googlemaps and etc. Google’s original Hbase hapdoop data store.
Google BigQuery : Web interface to scan and query millions, billions, trillions of rows.
Google DataProc : Managed Hadoop, Spark Pig, Hive combined into one interface. Spin up a multiple cluster nodes in seconds and run a spark, pig, hive job using google’s compute power.
(better than amazon emr)
Summary: pay-as-you-go for cloud computing service. Watch out Amazon Web Services and Azure.
google data studio : neat visualization app for data.