IoT - cloud computing &
analytics
1
2
IoT – transformation
Connected
Why Build an IoT?
3
IoT Analytics Architecture
Device Management
Google Cloud Platform
Amazon Web services
Devices
4
IoT Architecture
Gateway
Devices
5
IoT Architecture
Connectivity
Gateway
Devices
6
IoT Architecture
Analytics
Storage
Connectivity The value of IoT data for organizations
Better human productivity
Gateway
Improved equipment maintenance
Devices
Operations optimization and automation
7
• The data generated from IoT devices
turns out to be of value only if it gets
subjected to analysis, which brings
data analytics into the picture.
• Data Analytics (DA) is defined as a
process, which is used to examine big
and small data sets with varying data
properties to extract meaningful
conclusions and actionable insights.
• These conclusions are usually in the
form of trends, patterns, and statistics
that aid business organizations in
proactively engaging with data to
implement effective decision-making
processes.
8
• Streaming Analytics: This form of data analytics is also
referred as event stream processing and it analyzes
huge in-motion data sets.
• financial transactions, air fleet tracking, traffic analysis etc. can
benefit from this method.
• Spatial Analytics: This is the data analytics method that
is used to analyze geographic patterns to determine the
spatial relationship between the physical objects.
• Location-based IoT applications, such as smart parking
applications can benefit from this form of data analytics.
• Time Series Analytics: As the name suggests, this form
of data analytics is based upon the time-based data
which is analyzed to reveal associated trends and
patterns.
• weather forecasting applications and health monitoring
systems can benefit from this form of data analytics method.
9
Reasons why IoT data requires analytics
• Volume
• Structure
• Life-saving situations
• Driving revenue
10
IoT Analytics
• Internet of Things (IoT) analytics enables organizations
to leverage the massive amounts of data generated by
IoT devices, using analytics stacks.
• IoT analytics is often considered a subset of big data,
involved with combining heterogeneous streams and
transforming them into consistent and accurate
insights.
• it is often complex to integrate the many types of IoT
devices with existing ecosystems and analytics tools.
• This is why, organizations deployed Industrial IoT (IIoT), a
technology built for collecting and analyzing data from
sensors.
11
Types of Data Analytics
12
Use Case – Air Quality measurement
• https://www.iqair.com/us/india/tamil-nadu/vellore
Source: national air quality index, CPC data
Visualization created using a thinkspeak
13
Thingspeak
• ThingSpeak is a Web Service (REST API) that lets
you collect and store sensor data in the cloud
and develop Internet of Things applications
• Collect data from sensors “things”
• Visualize them instantly
• Act on the data – e.g when the temp goes
beyond 340 C, then give a tweet.
14
Use Case – Air Quality measurement
Diagnostic capabilities are
typically extensions to
dashboards that permit
humidity
users to drill into data,
compare it, and visualize
correlations and trends in
an ad-hoc manner.
temperature 15
Use Case – Air Quality measurement
16
Process – Predictive analytics
Data Data Performance
Cleaning Modelling analysis
17
Data
Cleaning
• With millions of connected devices, a typical data processing
platform might be required to deal with billions of such incoming
events every day.
• Even though processing this amount of data is obviously a
considerable and by no means trivial technological challenge, it is
clear that the device data itself – even when stored in a
preprocessed form – is not something actionable.
• To get actionable insights, the collected data must be analyzed.
• One type of task that can be effectively tackled with data analysis
in the IoT is anomaly detection.
• Its goal is to identify unusual behavior in connected devices that
differs significantly from what has been observed before or from
what is expected.
18
Data
Cleaning
• Remove Unwanted Observations
• Check for Irrelevant Observations
• Filter Out Outliers
• Handle Missing Data
• Convert Data Types
19
Data Predictive Analytics Models
Modelling
• Classification Model
• It puts data in categories based on
what it learns from historical data.
• Classification models are best to
answer yes or no questions, providing
broad analysis that’s helpful for
guiding decisive action.
• For an online banking provider, “Is this
a fraudulent transaction?”
• For a data from a fitbit tracker, “Is this
heart pulse rate leads to heart attack
or not”
• Any machine learning or a deep
learning model which requires a
training and learning fits into this. 20
Data
Modelling
• Clustering Model
• The clustering model sorts data into separate,
nested smart groups based on similar
attributes.
• Using the clustering model, they can quickly
separate customers into similar groups based
on common characteristics and devise
strategies for each group at a larger scale.
• grouping loan applicants into “smart buckets”
based on loan attributes
• identifying areas in a city with a high volume
of crime using the data from surveillance
system
• benchmarking SaaS customer data into
groups to identify global patterns of use.
21
Data
Modelling
• Forecast Model
• Deals in metric value prediction, estimating numeric value
for new data based on learnings from historical data.
• A call center can predict how many support calls they will
receive per hour.
• The forecast model also considers multiple input
parameters. If a restaurant owner wants to predict the
number of customers she is likely to receive in the
following week, the model will take into account factors
that could impact this, such as:
• Is there an event close by?
• What is the weather forecast?
• Is there an illness going around?
22
23
24
Cloud Interface
Interface With Cloud
Channel Settings
API Keys, mainly Write API
1.
1. Connect the NodeMCU with micro USB cable to your PC.
2. Press and hold flash button and press the reset button once
and now you leave the flash button.
3. Select the NodeMCU v1.0 as board.
4 Select 115200 as baud rate and select the correct COM
port.
5 Press upload button, it may take 2 min to compile the code
and another one minute to upload to NodeMCU.
• #include <DHT.h> // Including library for dht
•
• #include <ESP8266WiFi.h>
•
• String apiKey = "xxxxxxxxxxx"; // Enter your Write API key from ThingSpeak
•
• const char *ssid = "xxxxxxxxxx"; // replace with your wifi ssid and wpa2 key
• const char *pass = "YYYYY";
• const char* server = "api.thingspeak.com";
•
• #define DHTPIN 0 //pin where the dht11 is connected
•
• DHT dht(DHTPIN, DHT11);
•
• WiFiClient client;
• void setup()
• {
• Serial.begin(115200);
• delay(10);
• dht.begin();
•
• Serial.println("Connecting to ");
• Serial.println(ssid);
•
•
• WiFi.begin(ssid, pass);
•
• while (WiFi.status() != WL_CONNECTED)
• {
• delay(500);
• Serial.print(".");
• }
• Serial.println("");
• Serial.println("WiFi connected");
•
• }
•
Case Study - Open Source e-Health sensor
platform
• E-Health is an open-source sensor platform that
offers a wide range of bio-signal sensing capabilities
to open-source hardware platforms.
Analytics • Pulse sensor
• Oxygen in blood sensor
Storage • Body temperature sensor
• ECG sensor
Connectivity • Galvanic Skin Response (GSR) sensor
• Glucometer
Devices
• Accelerometer
• Blood pressure sensor
• EMG sensor
33
• The ports and pins on the shield are designed specifically for the
sensors listed above.
• All sensors are noninvasive and are designed with electrodes, clips
and bands among several other features that make it easy to attach
them safely onto the body.
• Using these sensors one can monitor the real-time state of a patient
and be able to make a medical diagnosis.
34
• Data that is collected from the patient can be stored remotely. This is
done through Wi-Fi, Bluetooth, GPRS, 3G, ZigBee or 802.15.4
depending on the application.
• The data is always secure regardless of which mode of
communication you use.
• This is because the platform has several security levels. The
communication link layer has AES128 for ZigBee/802.15.4 and WPA2
for Wi-Fi, while the application layer uses HTTPS (secure) protocol.
• So, logged data is always secure.
35
• Pulse sensor
• Heart rate data - designing an exercise routine,
studying your activity or anxiety levels or just want
your shirt to blink with your heart beat.
• The Pulse Sensor Amped is a plug-and-play heart-
rate sensor for Arduino.
• Simply clip the Pulse Sensor to your earlobe or
finger tip and plug it into your 3 *or *5 Volt Arduino
Pulse wave is the change in the volume of a blood vessel
that occurs when the heart pumps blood, and a detector
that monitors this volume change is called a pulse
sensor.
36
• Oxygen in blood sensor
• Pulse oximetry reading, a small clamp-like device is placed on a finger,
earlobe, or toe.
• Small beams of light pass through the blood in the finger, measuring
the amount of oxygen.
• It does this by measuring changes of light absorption in oxygenated or
deoxygenated blood.
37
• How data analytics can be applied from Pulse
oximetry reading
• whether someone needs help breathing
• evaluate how helpful a ventilator is
• monitor oxygen levels during or after surgical procedures
that require sedation
• assess someone’s ability to tolerate increased physical
activity
• evaluate whether someone momentarily stops breathing
while sleeping
38
Be Close Elderly monitoring
• Home monitoring systems for the elderly, allowing them to stay at
home even when they may be fragile or lack motility.
• The system consists of a number of motion sensors as well as a
weight pad (to see when they’re sleeping) and a panic button that
notifies the authorities and kin if there’s something wrong.
• Data analytics
• Get an alert if the medicine cabinet hasn’t been opened by a specified time
every day.
• Know instantly if an exterior door is opened at night, or at an unusual time.
• Know if the elder at home isn't moving around their home during the
daytime.
39
• Fall detection - Alert you if a probable fall occurs
• An alert is triggered when it detects that person has not
moved from high-risk fall areas (such as a stairwell or
restroom) in an unusually long time.
• Motion sensors can work in combination with Contact
sensors affixed to exterior doors/stairs to detect when your
inmate has left that place.
40
