Hi, my name is Zoe Gupta and I'm a data scientist working on the T team. I'm here today to answer the question. What is graph based data science? And how is it different from traditional data science? Let us begin by defining what data science actually is. Data science simply put is where data scientists use data to answer questions. It is an interdisciplinary field that uses various scientific processes.
There are multiple ways to do this data, scientists could use psychic learn to build a machine learning model or perhaps do some basic exploratory data analysis using pandas or even construct simple visualizations using the data to gain knowledge or insights. Usually data scientists use flat or tabular data for their analysis. However, in the case of graph data science, the relationships are used to answer questions.
Relations between data are an important source to get answers. Like we have a set of different methods or a tool kit in data science. We have different methods of finding answers. In graph based data science. We can visualize the data in the form of a graph and visually inspect the graph to see what we can find or we can query the graph asking very specific questions similar to SQL or we can use graph algorithms spanning the whole graph to get some answers.
Now, looking at an overall picture graph analytics approaches fit across a spectrum of local patterns based approaches to global computations by this. I mean, if we know exactly what we are looking for and we are focusing on a narrower part of the graph, we can use a query to get the answer we're looking for this query helps us make a decision based on accessing the local part of the graph.
We're trying to identify patterns in the graph in a very focused manner. On the other hand, graph algorithms are used for an overall global analysis of the graph. Here, we are trying to learn the overall network structure versus answering a specific question. Hope this was useful.
Thank you.
