2a:null 2b:["$","title",null,{"children":"Why I choose Data Journalism? | Shuai Hao | Observable"}] 30:T425,md`My father buys Chinese National Geography every month. Since I was six years old, he and I have read this magazine, which is famous for its exquisite illustrations, and similar to National Geographic, together. As a six-year-old boy, I intuitively understood that dense, text-heavy articles were dull, while stories with pictures and fewer words were much easier to understand. As I got older, I continued to see images and visual aids as extremely vital to storytelling, because they can convey meaning much more powerfully than words alone. Through formal training and experience in journalism and mass communications, I have learned to communicate using visualizations that bring the nuance of data to life through the impact of pictures. I am now in Columbia’s dual degree program to further explore the possibilities of data-fueled visual innovation in storytelling.

_My father’s Chinese National Geography collections_`31:T1295,md`I made multiple data related projects at Columbia. In one, I designed a database system in PostgreSQL using NYC open data to display crimes, monuments, eateries, hiking trails and other parameters in all public parks at New York City. I built a user interface in Python and JavaScript for visitors to access the database, which involved add, edit, delete and search functions. _I designed a NYC park database in PostgreSQL and Python with my classmate_

In another, I designed [an interactive website](https://haoshuai999.github.io/) in D3.js and scrollama.js to explore why and how terrorism evolved over the years by analyzing 180,000 rows of data about weapons, victims and perpetrators of terrorism attacks in the last 50 years using pandas.

In other smaller data projects, I used pandas to clean the dataset, applied linear regression or logistic regression model on the data and made preliminary visualizations in matplotlib or R library ggplot2. In terms of data gathering, I filed a FOIL request and got the data about mobile food vendors in New York City.

The following are some Python code for analyzing a dataset about initial coin offering:
~~~~ import csv import numpy as np import matplotlib.pyplot as plt import statistics # The dataset is about the size of initial coin offering (ICOs) from February 2014 to July 2018 # The dataset also includes the name and close date of ICOs, which can be used for further analysis # The following script provides the median, quartiles, the mean and standard deviation of ICO size. # Next step: A company called EOS raised $4,200 million dollar through ICO on June 1, 2018 # The second largest ICO size is only about $850 million dollar up to July 2018 # How EOS raised such amount of money is worth investigating # And Syed Shoeb, a writer for Hacker Noon, wrote an article on medium about the possibility of scam for EOS's crowdfunding (https://hackernoon.com/is-eos-a-scam-eos-ico-suffers-collusion-allegations-complete-analysis-e21e68a2a5c7) # Values are specified in million dollars # Data source: Coindesk ICO Tracker (https://www.coindesk.com/ico-tracker) # Open the csv file about ICO data data_file = open("ICO Data Export.csv","r") # Convert to DictReader data_reader = csv.DictReader(data_file) # Set a list for total ICO size total_ico_size = [] # Clean the data by removing dollar sign, dash and comma for row in data_reader: total_ico_size.append(float(row[data_reader.fieldnames[2]].replace("$","").replace("-","0").replace(",",""))) # Calculate and print the quartiles, interquartile range, upper and lower bound quartiles = np.percentile(total_ico_size, [25,50,75]) iqr = quartiles[2] - quartiles[0] k = 1.5 upper_bound = quartiles[2] + k * iqr lower_bound = quartiles[0] - k * iqr print("Median = %d" % quartiles[1]) print("First quartile = %d" % quartiles[0]) print("Third quartile = %d" % quartiles[2]) print("Upper bound = %f" % upper_bound) print("Lower bound = %f" % lower_bound) print('\n') # Calculate mean and standard deviation mean = statistics.mean(total_ico_size) sd = np.std(total_ico_size) # Store and print the mean, 1, 2 and 3 standard deviation in a list called mean_measure mean_measure = [mean] print("Mean = %f" % mean) for i in range(4): lower_sd = mean - (i * sd) upper_sd = mean + (i * sd) mean_measure.append(upper_sd) mean_measure.append(lower_sd) if i > 0: print("%d standard deviation = [%f , %f]" % (i, lower_sd, upper_sd)) # Plot the histogram and exclude the outlier $4,200 million dollar to make the histogram easy to read # But the outlier is not excluded when calculating mean, median and other statistical variables histogram = plt.hist(total_ico_size, bins=100, range=(0,850), histtype='bar', color= 'blue') # Plot the quartiles, upper bound and lower bound plt.vlines(quartiles, 0, 450, colors= 'red', linestyles='dashed', linewidth= 1) plt.vlines(upper_bound, 0, 450, colors= 'red', linestyles='dashed', linewidth= 1) plt.vlines(lower_bound, 0, 450, colors= 'red', linestyles='dashed', linewidth= 1) # Plot the mean, 1, 2 and 3 standard deviation plt.vlines(mean_measure, 0, 450, colors= 'yellow', linestyles='dashed', linewidth= 1) plt.show() ~~~~
I visualized datasets about NBA team performances with Tableau and felonies in NYC with D3.js and built a local server showing astronomy pictures from NASA in Flask and jQuery. The data visualization, user interface design and database courses I’m taking this spring will further enhance my data manipulation and visualization skills. `32:[] 2e:{"id":"96bbe3ad17b25442","slug":"why-i-choose-data-journalism","title":"Why I choose Data Journalism?","thumbnail":"3ebd3689677e682c5cb3c2f84098736ee4e78a36e76a1da1a72a5a876988694b","access_level":"public","create_time":"$D2019-06-16T01:22:50.530Z","update_time":"$D2019-06-16T01:22:50.530Z","view_time":null,"first_public_time":"$D2019-06-16T01:22:57.812Z","fork":{"id":"754c87f221f7b18a","slug":"why-i-choose-data-journalism","title":"Why I choose Data Journalism?","owner":{"id":"4d5e44fed9160cc9","login":"haoshuai999","name":"Shuai Hao"}},"stars":0,"starred":false,"owner":{"type":"team","id":"1eea9d58f7a2b820","login":"shuaihaofzny","name":"Shuai Hao","avatar_url":"https://avatars-next.observableusercontent.com/69c48036e8f4ab69047625ac1187d4d9926cd76ff47c9467416be9d448b00855","tier":"starter_2024"},"version":77,"publish_version":77,"role":null,"authored":false,"authors":[{"type":"individual","id":"a4290be9b862ad07","login":"shuaihaofzny","name":"Shuai Hao","avatar_url":"https://avatars-next.observableusercontent.com/69c48036e8f4ab69047625ac1187d4d9926cd76ff47c9467416be9d448b00855","tier":null}],"members":[{"type":"individual","id":"a4290be9b862ad07","login":"shuaihaofzny","name":"Shuai Hao","avatar_url":"https://avatars-next.observableusercontent.com/69c48036e8f4ab69047625ac1187d4d9926cd76ff47c9467416be9d448b00855","tier":null,"role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"subscribed":null,"forks":0,"collection":null,"type":"notebook","body":{"title":"Why I choose Data Journalism?","cells":[{"id":0,"value":"md`# Why I choose Data Journalism?`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":8,"value":"$30","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":20,"value":"md`Since I grew up reading magazines that used both artistic expression and scientific analysis to tell thrilling and fascinating stories about the world, I was deeply curious about both. In my freshman year at Communication University of China, I was a student in the school of computer science. I learnt a lot of courses about computer science, such as Advanced Mathematics, Programming Methodology, Data Structure. Programming opened a new door for my future. I wrote programs to work out math problems, practice basic operations of database and built [a simulation system about book management](https://github.com/haoshuai999/Library-Management).`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":23,"value":"md`In the spare time, my interest is developing interesting and attracting websites in the college laboratory. I was curious about expressing information through creative ways, like what data journalism students did in their interactive works. Good score in my freshman year offered me the chance to switch major from information security to journalism, the best major in the university. However, the gap between art and science made me hard to adjust. I was neither willing to give up knowledge about journalism, nor happy to stop learning programming courses for my future career.\n

\n_The interactive map I designed for 2017 Data Journalism Competition of China_`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":25,"value":"md`During the last three years of college, I have explored programming, data journalism, and visualization through self-study, while maintaining a superb academic record. In my sophomore year, I learnt Python to scrap data from the web, and Tableau to present graphs online and in school newspaper. I also participated in massive online open courses on global data journalism, and created interactive projects for a data journalism competition. When visualizing those projects, I was always the one responsible for technical part.\n\n

\n_I was the chief editor of the school newspaper_`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":27,"value":"md`In the summer of 2017, I joined Caixin Media, one of the most well-known finance media companies in China, as a data journalism intern. I learned many valuable lessons about data analysis and visualization from my colleagues, and published a number of news articles on topics as diverse as finance, politics, education, and science. Most notably, one article I wrote about suicide rates among older adults received 700,000 views, and was featured on the homepage for five days, which got the attention of some of the company’s senior executives. `","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":29,"value":"md`My work at Caixin Vislab, a data journalism award winning team, was both challenging and fulfilling. However, I also noticed some drawbacks to working in data journalism in China more broadly. Most Chinese reporters know little about programming. Most Chinese journalists are self-taught when it comes to the more technologically advanced parts of their jobs. Although I have a good foundation in both programming and web design, I still need more systematic training to prepare myself to enter an industry where new media knowledge and training are scarce and sorely lacking.\n

\n_Caixin Vislab won 2018 Best Data Journalism Team in two consecutive years_`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":31,"value":"md`Columbia University is the best university in journalism and engineering education, which is regarded as the ideal school by data journalists around the world. I took courses of fantastic professors in both schools and each of them has strong experience in journalism or programming. The courses about Computational Journalism and Computer Science is vital to my technical and reporting skills.`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":33,"value":"$31","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":37,"value":"md`
Nowadays, I am confident that my robust interdisciplinary academic and professional experience has fully prepared me for the unique challenges and opportunities I will encounter through my future career as a data journalist.`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"}],"files":"$W32","stdlib":"1","titling":"auto"}} 2f:"$2e:body"