Yifeng LIU

Abstract

With a undergraduate background in architecture and statistics, I have grown a strong interest in data-driven urban decision-making, applying the particularly interested in developing computational models and algorithms to support city planning processes. I’m a customer based person, and moreover, I believe that in this era of personal activity data explosion, it is necessary for a planner to participate in data science, so that the behavior of profiting from the public can be turned into the profiting for the public.

In my undergraduate years, I have researched in serious of project in mining the urban infrastructure or marketing via different methods across traditional CV, segmentation and trajectories mining. Describe my project about running trajectories here. Describe my project about housing plan here. Description of the project can be found in the following section.

Since I started my graduate studies at MIT, I start to dive into methodology and algorithm development studies. I have learned courses across optimization, machine learning, and transportation simulation(reinforcement learning). FLP Julia Package here. Running Behavior Analysis here. Now is density prediction, but wanna develop it to a Sequence data prediction Reinforcement Leaning here I found there is serious of works we can do in this field,

But now I am confused in my own decisions: If I am interested in the impact of computational algorithm research on urban construction, which path should I choose. How can I explore my career path and apply operations research or large models to real industries? Should I pursue a PhD? Because the relevant research-level positions in the United States require a PhD degree for internship; Or should I go straight back to the Chinese for industrial practice? I find that because of China’s social system and cultural background, we have more opportunities for industrial practices, whether it is because of the denser population in super cities, or because of the richer life in urban centers, such as food delivery, delivery and Internet economy. Or should I start a business? I now have a team doing some small trials with sandbox support, and I’m back home for spring break talking to alumni investors and I’ve got a lot of verbal commitments.

I am more confused now because I am not familiar with the industrial environment in the United States.

Research & Design Experiences

Allocating Covid Testing Stations in large-scale

Covid testing stations allocating optimization and decomposition in 2022 Lockdown Shanghai.

Jogging Analysis

Analyzing jogging activity patterns and adaptation to public health regulation.

Beijing Infrastructure

Analyzing local virbrancy, running activity, and real estate prices in Beijing.

Streets and Real Estate

Investigating the impact of perceived micro-level neighborhood characteristics on housing prices in Shanghai

Street Computer Vision

Research on citizen streetscape evaluation based on crowd sourcing survey and the application of navigation optimization.

Informal-Mapping

Computing places and human activity in data-absent informal urban settlements.

This file is the brainstorming draft of PathBy project.

the idea start from a previours project Portal

Business Plan Draft 1.0

In short: trafic path based facility recommendation system

imagine a world that you don’t need to travel away to a crowded restaurant, but the dilicious will suddenly appear around you. new events!! new restaurants!! new friends!! new opportunities!! new life!!

It’s a new Yelp, not recommend you to visit a place, but bring the diliciours around you.

it’s like a new Uber, not for taxi, but for facilities.

diagrams

Problem

even if people live in city for years, travel through the city for 2 hours per day, they will only know about 3 groups of facilities:

1. facilities around their home
2. facilities around their work place
3. facilities recommended by Ins and Yelp.

we can see that these three group pointer to 3 different kinds of urban functions: residential, work,and commercial. Usually, if someone want to go hangout with friends, he/she will go to the facilities around his/her home, or the facilities recommended by Ins and Yelp. the first one is limitated area, and the second one will cause huge traffic load.

The way people use facilities and the spatial correlation is weak, resulting in a lot of extra traffic demand and facility load on the basis of necessary traffic.

Maybe its a very Megacity problem? some more Asin area studies will help.

Solution

product 1: metro advertisement system

at the first stage, due to lake of personal data, we will focus on the metro advertisement system: with metro system and POI around metro stations, we can build the system, and will recommend facilities to people based on their OD trafic path.

product 2: personal facility recommendation application

as more and more people use our product, we will have more and more personal data, and we can build a personal facility recommendation application based on individual interest and daily path: recommend your future dinner plan!

based algorithm

recommend facilities to people based on OD trafic path

objective function:

Local index (need research support): extra distance consumption is the shortest; The highest recommended cumulative index

Overall indicator: Highest utilization rate of facilities, lowest traffic load.

Company purpose

Based on people’s trafic partten in Megacities, recommend facilities to them.

to customer:

With this product, we will help people to save time and energy on finding facilities, and help facilities to get more customers.

to city:

• A soft way of facility assolotation - for RealEstate economy;
• reduce the traffic load caused by - people’s extra trafic demand for far away facilities;

to commercial facilities: a new way of advertisement! more chance and more space for your new opportunities!

Market Size

(based on the research on ‘Problem’ part , show the portential martket and its influence)

Competition

Google map(Gaode Map)

Yelp(Meituan)

Business Model

…

Team

Yifeng

MIT DUSP MCP student 25’

2023, Gratudted from Tsinghua University, School of Architecture.

Urban informatics researcher. Research interest on facility management and Currently work on a project about the local optimization of facaility allocation problem. with wild experience about the inforstructure distribution analysis.

Ceci

Harvard GSD Mdes-public 25’

[education bg]

[short bio]

I’m looking for:

engineer:

• who have strong skillset in transporation system analysis, and optimization. I can do that but im good engough to do that alone.
• have skill set in spatial database or distributed computing
  • ML skillset of course, especially DL and large scale

CFO: no reason.

some questions/todo I have now:

• which city should I start with in demo?

• should I builid this project with academic research (show its social impact and more science based optimization algoroithm development) or just a startup (demo or die)?

• Singepore is too small for our project? where should be the market research base?

• market research
• prototype and test with users
• business model

##