23
Big Data Big Data ecosystem ecosystem

Big Data Ecosystem - 1000 Simulated Drones

Embed Size (px)

Citation preview

Page 1: Big Data Ecosystem - 1000 Simulated Drones

Big DataBig Dataecosystemecosystem

Page 2: Big Data Ecosystem - 1000 Simulated Drones

Presentation about Espeo's BigData ecosystem based on 1000simulated drones flying around

Poznan city

Page 3: Big Data Ecosystem - 1000 Simulated Drones

Drones produce and collect realtime sample data like:

latitude,longitude,height(m),temp(C),wind(m/s),humidity,air-polution

Page 4: Big Data Ecosystem - 1000 Simulated Drones
Page 5: Big Data Ecosystem - 1000 Simulated Drones

Drone soft written in Scala language

Page 6: Big Data Ecosystem - 1000 Simulated Drones

Drone soft written in Scala languageIn real time it streams data to the server

Page 7: Big Data Ecosystem - 1000 Simulated Drones

Drone soft written in Scala languageIn real time it streams data to the serverUsing Kafka

Page 8: Big Data Ecosystem - 1000 Simulated Drones

Drone soft written in Scala languageIn real time it streams data to the serverUsing Kafka

Page 9: Big Data Ecosystem - 1000 Simulated Drones

On a server, data is read by Spark Streaming.

It allows us to:

Page 10: Big Data Ecosystem - 1000 Simulated Drones

It allows us to:

save data to Cassandrasend calculated data to browser through websocketsend it to another Kafka consumersave the whole log to Hadoop cluster

On a server, data is read by Spark Streaming.

Page 11: Big Data Ecosystem - 1000 Simulated Drones

By saving logs to Hadoop cluster, we can lateraccess those logs, if we didn't save somethingin Cassandra

Page 12: Big Data Ecosystem - 1000 Simulated Drones

By sending data to the browser throughwebsocket, we can see where our drones are inrealtime, monitor sensors and much more

Page 13: Big Data Ecosystem - 1000 Simulated Drones
Page 14: Big Data Ecosystem - 1000 Simulated Drones

By using Cassandra and Apache Spark datascientists can analyze given data later,by using:

1. Apache Zeppelin - Apache Spark(df, RDD) + Scala - Apache Spark MLLib2. Azure Machine Learning

Page 15: Big Data Ecosystem - 1000 Simulated Drones

We prefer to use Azure Machine Learninginstead Spark MLLib because it is much easierto understand - and design new predictions

Read our blog post about Azure ML:http://espeo.eu/blog/azure-machine-learning-predictions/

Page 16: Big Data Ecosystem - 1000 Simulated Drones

CompleteCompleteecosystem diagramecosystem diagram

Page 17: Big Data Ecosystem - 1000 Simulated Drones

Drones

Wifi

Page 18: Big Data Ecosystem - 1000 Simulated Drones

Drones

Wifi

Page 19: Big Data Ecosystem - 1000 Simulated Drones

Drones

Wifi

websocket

Page 20: Big Data Ecosystem - 1000 Simulated Drones

Drones

Wifi

websocket

Page 21: Big Data Ecosystem - 1000 Simulated Drones

Drones

Wifi

websocket

Page 22: Big Data Ecosystem - 1000 Simulated Drones

Drones

Wifi

websocket

API

Page 23: Big Data Ecosystem - 1000 Simulated Drones

Drones

Wifi

websocket

API