zy16373
Nov 25, 2017
Graduate / SOP for Data Science Admission; What exactly is a hacker? [4]
Hi all, I've taken your advice and modified my essay accordingly, could you all please check the 2nd version and give some more advice, thanks a lot!
My strong interest to solve problems with big data was triggered by one data science lecture I attended at XXX University. The professor described how he used trajectory data to analyze the behavior of stowaways from Mexica to the U.S. Such first insight of how data scientists approach real-world problems incited me to delve further into the field.
The course also introduced me to fundamental data science knowledge, including database design, social network analysis as well as statistical tools. I programmed in R to analyze data correlations, generated linear regression models to predict students' GPA scores, based on their GRE scores. One project during the course that I found particularly interesting was looking at data from social media users to find the relationships between followers and the people they follow. By identifying an influencer in a healthcare forum with PageRank and centralities, companies could deliver healthcare information and related promotions to certain social circles more efficiently.
After returning to college after my visit to Korea, I kept on thinking of ways to analyze social network data and make the most of them. In addition to the field of healthcare, I believe that this technology could benefit many other disciplines, such as business and political elections. So, I started another research project at the Nvidia-Joint lab, where I used CUDA parallel programming to accelerate network centrality calculations. I boosted the efficiency by paralyzing the centrality calculations and using a directed graph with more than 1,000 nodes and edges, the total computation time was reduced from 192 to 2 ms. The power of GPU calculation enabled massive data processing, which can be used to achieve real-time network centrality calculations. In this way, a politician could identify opinion leaders, as well as their attitudes to politics, and adjust his election strategies accordingly.
Being majored in Computer Science, my machine learning experiences enable me to analyze data efficiently and creatively. In the course of a research project I completed in the summer of 2016, I trained a classifier to identify obstacles on roads - a useful technique for unmanned vehicles, using the Nvidia deep learning framework DIGITS. My team labeled the pictures to be tested and then trained AlexNet to identify them. However, due to a number of mistaken configurations, the system's performance was not ideal. After reading papers discussing deep learning networks, I enlarged the dataset and adjusted the learning rate and batch size, and, with these adjustments, I boosted the accuracy of the model to over 80%.
My resolve to specialize in data science is strengthened by my current work of final year dissertation. My project examines how the analysis of the patterns in player trajectories can be improved using machine learning techniques. Namely, while the new availability of player-tracking data facilitates detailed analyses of the patterns in player movement, traditional tactical analysis largely relies on the knowledge and manual labor of domain experts, which is a very expensive and unscalable approach. Being a huge basketball fan, I have wondered why I am unable to identify frequently appearing patterns in professional basketball matches, which describe teams' tactics and behaviors without manually labeling all offensive tactics used. I plan to generate features that differentiate plays from one another, then use unsupervised machine learning techniques to extract frequently appearing player trajectories.
In this way, I hope to help teams to better understand how both they and their opponents play. Even though I am only an amateur basketball fan, by utilizing the power of machine learning, I believe we can discover the tactics used by professional plays as domain experts, making them even faster and more creative. The excitement of mining basketball plays motivates me to launch a career in sports data analytics, using the knowledge of data science to promotes the development of professional sports, particularly professional basketball.
I believe your program is the perfect choice to help me to achieve this goal for several reasons. First, I can deepen my understanding of statistical models, hone up skills of selecting right strategies for different problems. Further, the project-focused courses can provide chances using computational techniques to work through real-world challenges. I can also hone communication skills through group projects, which are essential when working in the industry. Ultimately, XXX's core value to promote the betterment of society, and further our understanding of the nature of the universe will guide me throughout my career as an ethical and creative data scientist. My creative and rigorous thinking showed in courses and research, my passion for sports, along with my strengthened understanding of data science through your program, will well-prepared me to provide unique insights into this industry.
Hi all, I've taken your advice and modified my essay accordingly, could you all please check the 2nd version and give some more advice, thanks a lot!
My strong interest to solve problems with big data was triggered by one data science lecture I attended at XXX University. The professor described how he used trajectory data to analyze the behavior of stowaways from Mexica to the U.S. Such first insight of how data scientists approach real-world problems incited me to delve further into the field.
The course also introduced me to fundamental data science knowledge, including database design, social network analysis as well as statistical tools. I programmed in R to analyze data correlations, generated linear regression models to predict students' GPA scores, based on their GRE scores. One project during the course that I found particularly interesting was looking at data from social media users to find the relationships between followers and the people they follow. By identifying an influencer in a healthcare forum with PageRank and centralities, companies could deliver healthcare information and related promotions to certain social circles more efficiently.
After returning to college after my visit to Korea, I kept on thinking of ways to analyze social network data and make the most of them. In addition to the field of healthcare, I believe that this technology could benefit many other disciplines, such as business and political elections. So, I started another research project at the Nvidia-Joint lab, where I used CUDA parallel programming to accelerate network centrality calculations. I boosted the efficiency by paralyzing the centrality calculations and using a directed graph with more than 1,000 nodes and edges, the total computation time was reduced from 192 to 2 ms. The power of GPU calculation enabled massive data processing, which can be used to achieve real-time network centrality calculations. In this way, a politician could identify opinion leaders, as well as their attitudes to politics, and adjust his election strategies accordingly.
Being majored in Computer Science, my machine learning experiences enable me to analyze data efficiently and creatively. In the course of a research project I completed in the summer of 2016, I trained a classifier to identify obstacles on roads - a useful technique for unmanned vehicles, using the Nvidia deep learning framework DIGITS. My team labeled the pictures to be tested and then trained AlexNet to identify them. However, due to a number of mistaken configurations, the system's performance was not ideal. After reading papers discussing deep learning networks, I enlarged the dataset and adjusted the learning rate and batch size, and, with these adjustments, I boosted the accuracy of the model to over 80%.
My resolve to specialize in data science is strengthened by my current work of final year dissertation. My project examines how the analysis of the patterns in player trajectories can be improved using machine learning techniques. Namely, while the new availability of player-tracking data facilitates detailed analyses of the patterns in player movement, traditional tactical analysis largely relies on the knowledge and manual labor of domain experts, which is a very expensive and unscalable approach. Being a huge basketball fan, I have wondered why I am unable to identify frequently appearing patterns in professional basketball matches, which describe teams' tactics and behaviors without manually labeling all offensive tactics used. I plan to generate features that differentiate plays from one another, then use unsupervised machine learning techniques to extract frequently appearing player trajectories.
In this way, I hope to help teams to better understand how both they and their opponents play. Even though I am only an amateur basketball fan, by utilizing the power of machine learning, I believe we can discover the tactics used by professional plays as domain experts, making them even faster and more creative. The excitement of mining basketball plays motivates me to launch a career in sports data analytics, using the knowledge of data science to promotes the development of professional sports, particularly professional basketball.
I believe your program is the perfect choice to help me to achieve this goal for several reasons. First, I can deepen my understanding of statistical models, hone up skills of selecting right strategies for different problems. Further, the project-focused courses can provide chances using computational techniques to work through real-world challenges. I can also hone communication skills through group projects, which are essential when working in the industry. Ultimately, XXX's core value to promote the betterment of society, and further our understanding of the nature of the universe will guide me throughout my career as an ethical and creative data scientist. My creative and rigorous thinking showed in courses and research, my passion for sports, along with my strengthened understanding of data science through your program, will well-prepared me to provide unique insights into this industry.