Stepping into the World of Data Science

Written by - Tejashri Nakhate, Anju Shakya and Hanny

Fluctuating Trends: An insight into data variability explored by the Data Science Club.

Data Science seems to be an intimidating field to step into. Understanding large datasets requires a combination of analytical thinking and practical application. It is basically the ability to interpret and derive meaningful insights from any given data, figure out the trends and patterns of real world data. This process involves not just theoretical knowledge but also a problem-solving mindset, curiosity, and creativity.

Data analysis is not just about numbers rather it is about exploring patterns, identifying trends, and making data-driven decisions. The best way to master this skill is by "playing with the data", which means experimenting with different datasets, visualizing information, and applying statistical techniques.

Working with data enhances logical reasoning and problem-solving skills. As you dive deeper into different datasets, you start recognizing patterns and this process sharpens your ability to think critically and apply data driven insights to real-world problems.

But many people struggle to read the data accurately due to the lack of statistical knowledge. They inaccurately determine causation and correlation between data due to similar trends in both the data. But in actuality further analysis and other factors must be considered to figure out the real relationship between two data. For instance, take an example of data that shows an increasing number of sunburns among people with ice creams, one might draw the conclusion that ice cream causes sunburn (even though we know that it is wrong). The real reason behind this is the common cause of hot weather, people staying out in hot weather would prefer eating ice cream to cool themselves, they would also have higher chances of sunburn due to being outside in hot sun. From this simple example we can see how easy it is to misinterpret data.

Data is not also not unbiased information. For example, if you use surveys to collect data then the sample chosen for the survey can be biased and lead to wrong data leading to faulty interpretations. One of the most harmful and long lasting of a sampling error was the study by Andrew Wakefield that falsely concluded MMR vaccines cause autism, this conclusion has now been debunked in thousands of better conducted studies. To learn and read more about sampling error check of this article - Sampling Bias and How to Avoid It

To begin your journey of data science lets see some of the most common formats used to store data and their specific use.

CSV (Comma-Separated Values): Used for structured tabular data.

JSON (JavaScript Object Notation): Used for web-based data and APIs.

XML (Extensible Markup Language): Common in web applications and configurations.

SQL Databases: Data stored in relational tables.

No SQL Databases: Data stored in flexible formats like documents or key-value pairs.

Log Files: Unstructured event-based records.

Once you get your data in any of the above formats you then import the data of your choice into software, could be R, Python or Excel. Then comes cleaning of data, organization of data, isolating the variables of interests and analyzing the trends and relations between the variables.  

With this introduction we welcome you into this data driven world.