Hi there! My name is Zhiqian Zhou. I am glad to meet you! I am currently a Master of Computer
Science student in University of Illinois at Urbana-Champaign. I am solid in
Python. I have experience in web development, machine
learning, and hope to explore more interested fields! I am fond of anime piano, and sports like
basketball, table tennic, etc. Feel free to contact me for possible future career opportunity and fun
things in anime piano ; )
Zhiqian Zhou
Software Developer
Timeline
Aug. 2015 - July 2017
Bachelor of Engineer in Software Engineer
I began my journey in Computer Science at South China University of
Technology. I studied C++ and Java and lay a solid
foundation. To explore more possibilities in my life, I join the 2+2 program and leave for
University of Miossri - Columbia.
Aug. 2017 - May 2019
Bachelor of Science in Computer Science
I continued to pursue my bachelor degree in University of Missouri -
Columbia. I joined in Prof. Dong Xu's lab and began my research in machine
learning field. Luckily, I have two papers published and one on submission now.
Aug. 2019 - Present
Master of Computer Science
I am so excited to receive an offer from University of Illinois at
Urbana-Champaign! I decided to become a full-stack software engineer and
lucikly received a summer intern offer from Facebook!
May-August 2020
Software Engineer Intern
I am honor to be a Software Engineer Intern at Facebook!
Be Part
Of My
Story!
Porjects
Smart Restrooms Monitor
Detection of Restroom Environment
Academic Search Engine
Search Engine with Visualization
Emojitrip Social Website
Travel Website with Emoji Feature
Publications
Dongpeng, L.*, Ming, L.*, Guangyu, S., Zhiqian, Z., Duolin, W., Jiaxin, L., Ryan, G.,
Jeffery, S., & Dong, X. A Deep Learning Model and a Mobile App to Assess Oil Spill for Environmental Water
Quality Monitoring. Manuscript in preparation.
Ming, L., Zhiqian, Z., Penghui, S., & Dong, X. (2019). Fuzzified Image Enhancement for
Deep Learning in Iris Recognition. IEEE Transactions on Fuzzy Systems. Advance online publication.
doi:10.1109/TFUZZ.2019.2912576
Built Smart Restroom using IoT devices to detect environment, sqlite3
to store data and React Framework for frontend website.
Date: September-December 2019
Responsibility: Frontend Development
Use humidity, VOC(volatile organic compound) sensors and switches (for restroom occupancy) to monitor
the envrionment of restroom. Connect sensors and switches with arduino like the picture above.
Configure ESP32 to facilitate host communication between Arduino and ESP32. The hardware directly post
data through ESP32 to backend, where the data will be stored in database for analysis. Frontend get
data from backend api.
Backend provides api for hardware to post and frontend to get. In addition to the instant data,
backend also calculate the average value for each kind of data and stored into the database. Frontend
will get history data and plot line chart for the data in the last 24 hours.
The frontend development include three parts. History data visualization, monitor for instant data,
and current weather and forecast. Users can check the monitor to know restroom occupancy, air quality
and history data. React Framework is used to update each section independently, where the instant data
is updated every 5 seconds and the weather and history data is updated one hour a time.
Academic Search Engine
Our academic search engine is designed to provide users with better
understanding of relationships between authors and advanced functions for paper. We implemented graph
visualization for author relationships to provide an overview for each author. In addition, we
developed multiple algorithms for users to choose for search engine, and built similar papers function
using LDA. Users can also view the topic distribution through a pie chart.
Date: August - December 2019
Responsibility: Full-Stack, Algorithm
Our Academic Search Engine provide user multiple options to choose their favorite algorithm when they
search. Currently we have five options for user to choose, BM25, BM25L, BM25+Rocchio, BM25L+Rocchio,
and Correlation. Moreover, we provided detailed visualization for both author profile and paper
detail. We put word cloud in our author's profile page to help users better understand the author. And
for each author, we have co-author graph visualization. More papers two authors have cooperated, the
bigger the node will be. By clicking the node in the graph, we can see each co-author's keyword to
understand this author's research interests and his co-author's research interests.
In addition to the author page, we also provide similar paper function for each paper using LDA
model. A pie chart is also generated for each paper according to the topic distribution in LDA model.
By clicking the View Details button in the page shown above, a visualization for LDA model will be
opened in another page.
Emojitrip Social Website
A social media web application that allows users to create emoji
reactions tagging on places of interests. Users can login, pick locations on the map, tag emojis and
like others’ reactions. They can also filter reactions or locations in multiple ways like physical
objects or emotions.
Date: August - December 2019
Responsibility: Full-Stack
Built emojitrip website with Django, Bootstrap Framework, Google API, sqlite3 and MongoDB Database. A
document stored in MongoDB contains location’s ID, a list of location types, and optional
recommendation implications. In the main page shown below, user can add new comment to any location or
click general search button to jump to the search page. In the search page, The search box can be
searched by location, user and emoji. The comment and the emoji on the map will be changed at the same
time.
By clicking the emoji on Google Map, a popup window will show the detailed information about that
location. Title is the location name, the second line is the top five popular emoji in this location,
and the third line is detailed address.
Moreover, recommendation function is provided using information theory in the popup window when user
click create comment button. In the search page, users can also use this function by clicking
recommend button. A stored procedure and trigger of scheduler in the backend is implemented to update
the recommendation data in MongoDB.
