With the connected devices emerging in every Domain of Industry day by day. We are entering into the smart world where the ways of doing things are being changed; with sensors collecting the huge data set to measure the physical world and taking Analytical step to take intelligent decision is becoming the new business model.
Traditional data analytics is incredible at clarifying information. You can produce reports or models of what occurred in the past or what’s going on today, adding valuable experiences to apply in organization, enable smarter decision making but Traditional Data Analytics are often static and of restricted use in tending to quick changing and unstructured information. With regards to IoT, it’s frequently important to distinguish connections between many sensor sources of info and outside components that are quickly creating many data points.
The core themes dimensions are –
- Internet of Things
- Data Science
- Architecture, e.g.: Hadoop, Azure, Spark, Storm
- Data pre-processing, Sensor fusion
- Algorithms/ Maths
- Programming (Python, SQL and Deep learning)
- Verticals(in our case IoT) but also extend to Telecoms, Supply Chain, Transport, Retail as applications of IoT
IoT Data Science has similarities but also some key differences. The differences arise primarily as we deal with multiple technology layers.
IoT applications also get complex with integration of embedded & connectivity technologies ( wireless sensor networks) – multiprotocols – security – data management & applications :
- Wireless Sensor Networks – Sensors acting as sources of data – Link layer protocols
- The data is largely time-series & structured .
- Need for Real time processing & streaming analytics
- Applications based Analytics – Data Science models
- Deep learning, Machines learning and multi-sensor data based modelling
- Security & Privacy- which are most critical – and with severe consequences if something goes wrong
Understanding Low-powered device (embedded micro-controllers) and sensor data acquisition. There are range of low powered micro-controller platforms which can easily prototype basic IoT applications e.g.: Arduino, Raspberry Pi, Beagleboard and more. Also interfacing external hardware device such as sensors, actuators with micro-controller platform. These interfacing hardware may comes as a standalone micro-chip unit,sensor breakout boards or sensor shields. Also working with real-time operating system concepts such as threading, processes scheduling, interrupts, I/O, memory management and timer would help to function the hardware platform in a more robust and efficient way.
For wireless communication between different machine one can use Bluetooth, Zigbee, Z wave or ipv6 protocol (TCP/IP protocol), routing , mesh network , or can use any OS which provides these features. CoAP, MQTT will also be needed.
Here you are having a gateway which is collecting data from different machines as well as sending data to cloud for further uses, like analytics, rule engine. Instead we do pre-processing of data at device edge and transfer the meaningful required context data to cloud. Many IoT applications like Fleet management, Smart grid, Twitter stream processing etc have unique analytics requirements based on both fast and large data streaming. It is critical to have a system that determines which data need to be processed immediately-at the edge-which data should be moved . (To be continued in Part 2)