Page 27 - Touhpad Ai
P. 27

For example, if you collected students' study hours and exam scores, data exploration might reveal:
                 u  Students who study more than 3 hours daily score above 80%.
                 u  A data error where a student scored 110 out of 100 marks.
                 u  Missing values where some students didn't enter their study hours.
                 These issues must be handled before training the AI model to ensure accuracy.
                 Data  exploration is  like  reading a story  hidden  inside  the numbers. It helps you
                 understand the quality of your data and prepares you to build a reliable AI model
                 later. Tools like Excel, Google Sheets, or programming languages like Python (with
                 libraries like Pandas) can help you summarise and visualise the data using charts
                 and graphs.















                                                                      Cleaning     Cleaned   Selection   Prepared
                                                                     Integration    Data    Transforming   Data

                 Let us understand the above 3 steps through a case study.

                                    Case Study: Understanding Mobile Phone Usage Habits Among Students
                  Objective: The purpose of this study is to analyse students’ smartphone usage patterns and determine their impact on
                  daily productivity and sleep habits. This case study demonstrates how data can be collected, explored, and interpreted
                  to make informed decisions—a key step in any AI or data analysis project.
                  Step 1   Identifying the Data

                  Before collecting data, clearly define the goal of the project. In this case, the goal is to understand how mobile phone
                  usage might influence productivity and sleep.
                  Next, identify the relevant features (variables) to collect:
                  u  Screen time (hrs/day): Numeric, Continuous
                  u  Time spent on social media (hrs/day): Numeric, Continuous
                  u  Number of phone unlocks (per day): Numeric, Discrete
                  u  Bedtime (average sleep time): Textual (can be converted to numeric
                     using 24-hour format)
                  u  Self-rated productivity (1 -5 scale): Ordinal
                  Optional additional variables for deeper insight: time spent on educational apps, number of notifications, or daily
                  study hours. Clearly identifying variables helps ensure the analysis is meaningful.
                  Step 2   Acquiring the Data

                  Once the variables are defined, gather the data:
                  u  Ask classmates to check their phone usage stats (available in smartphone settings) or estimate usage over the
                     past week.




                                                                                 Basic Concepts of Artificial Intelligence  25
   22   23   24   25   26   27   28   29   30   31   32