Part-1 – Data Regression
..IoT has changed our mindset from internet of ‘devices’ to ‘things’.
Wait.. this actually had already been happening behind the scenes through many years! From custom silicon chips, to OBC (on board computers), to Embedded systems (port based communication), SCADA (client server), to home/industry automation and now connecting every ‘thing’. The downsides and/or limitations of such a mesh network of every thing had been known for years too. Security, distribution, standardization to say the least, and money needed for physical enhancement of all the ‘things’ in the physical space.
IoT emerges as a business opportunity for many, and with advent of physical innovations ranging from better and faster silicon chips to Nano-technology, the future is exciting as we stand on the verge of next big thing...
For a business owner, the fact remains that ‘Internet’ of things, inherently needs physical enablement of ‘things’ to have multiple capabilities. For instance, whatever I need to connect to the IoT network bluntly requires me to compliment given physical object with necessary sensory hardware, communication engine and software, compliant to common protocols. Sure there are a few exceptions when one of these components may already exist or not needed at all. However today's things are hard-wired to the IoT enabling technology. Thus the fact remains that IoT has potential to increase capital investment, into assets which are not truly ‘retail’. In most cases, it will be worth, in some cases, it may not.
As a side note, many startups working in the space of IoT devices (physical) possess a risk of not being a true ‘off the shelf’ product, since there are not many shelfs they would have access to. Things may go positive as well, when a bigger brand acquires their technology and attempts to become a true ‘retail’ solution. Till such consolidation happens in IoT world, business owners cannot stop from taking advantage of the phenomenon
Smartphones, typically have all necessary ‘IoT’ components, and can be ‘smartly’ used as ‘IoT enablers’ where possible. Everybody has them, or they can be procured truly off the shelf. They are not attached to a given ‘thing’, benefiting in cases where certain things are to be connected to the network only for specific time span.
Smartphones come with their own specific challenges while attempting to produce reliable, consistent and repeatable sensor data. The peculiarities typically arise from the fact that smartphones come in variety of shapes, size, and quality. An accelerometer for instance, from a $10 phone may not always be as reliable as the one from a $500 phone. Differences arise in quality of data, frequency of data acquisition, consistency of data correctness or incorrectness, and calibration issues.
Certain statistical methods are known in the data researchers world, which are useful in ‘filling the gaps’ in various parameters of smartphone sensor datasets to make sure that given software can produce meaningful, reproducible data sets across multiple smartphone hardware sensors.
Taking accelerometer data as an example, following hands-on example explains how ‘Data Regression’ method can be used for enablement of smartphones as a reliable data acquisition device which can be attached to any ‘thing’
Data Regression Basics
Data regression is a technique for determining the statistical relationship between two or more variables where a change in a dependent variable (usually denoted by Y) is associated with, and depends on, a change in one or more independent variables (usually denoted by X).
How to enhance accelerometer values through data regression?
Simple Data Regression is used to examine the relationship between one dependent and one independent variable. After performing an analysis, the regression statistics can be used to predict the dependent variable when the independent variable is known.
Accelerometer gives acceleration values against time. While measuring acceleration we observed that the acceleration values were not recording at equal time intervals. Data regression can be utilized to evaluate relationship between acceleration and time.
In this case there is only one dependent and independent variable so we can use simple data regression.
Let’s try out Data Regression using a simple tool available on most of the computers, Microsoft Excel. Following is a step by step guide to experience power of data regression using the spreadsheet program.
Step 1: Open the Excel program. Copy and paste the data which we have to use for data regression into columns A and B beginning in row 1 of a blank worksheet.
Step 2: Select the Data ribbon menu, then Data Analysis command on the Analysis tab. A popup box will appear. Scroll down and select Regression. Click OK.
Step 3: The Regression wizard will be displayed. In the Input Y Range field, select all the values in the acceleration column of the data table
Step 4: In the Input X Range field, select time values from time column of data table same manner as Step 3.
Step 5: In the Output Range field, select cell E1 so that the regression output will begin in that cell
Step 6: Verify you have no checks in all remaining boxes. Your wizard should be identical to the graphic below:
Step 7: Click OK to 'run' the regression. Verify your output appears as follows
Step 8: Write the cost formula based on the regression output as:
Y = -0.000384 x + 0.136045431
The acceleration is 0.136045431 plus -0.000384 times the number of time.
Step 9: The acceleration value expected if 185-time value is:
Y = (-0.000384 * 185) + 0.136045431 = 0.06500543