 I would like to invite our next speaker, Dr. Nong Li, who is the Deputy Director in the Innovation Center for Technology, Beijing Jinghua, Tonghe Urban Planning, and the Design Institute. And before this, he has also spent seven years as a senior engineer in China tackling our urban planning and the design. And in recent years, he talked about the trend of data-driven planning, testing new data, and the tools for various issues in cities, transitioning the planning from a static process into more of a dynamic process. Okay. Now, I'm doing a company with some data analyzed, very intensive data analyzed. So my, the environmental problem, especially the waste problem is not my major work. So I give some relevant project to show you what we are doing right now, and something maybe you'll get some inspiration. Okay. And some introduction. My company is affiliated with this Jinghua University. It's purely urban planning and design institute with more than 1,000 people. So my innovation center is about 50 people. We have half of the camera and half of the data scientists, our coder. So we're working together to get some analysts to support a better plan and a better management. So this major six kind of work we are doing for different cities in China, like big data consulting, fusion analysis, monitoring, or some points and valuation kinds of things. It's not just focused on the waste problems about the population, industry, a little bit of comprehensive issues in Chinese cities. Chinese term, we call it Cheng Shui, okay. City disorders, it's not a term in English. I communicate with some overseas professions. They do not say this word in English word. I don't know. Like we call it city disorders in China is popular in Chinese policy texts, like some urban sprawl, traffic, and landmines kind of things. And when this UN project is more balanced for the whole society, for all human beings. And I think the planet, the term like urban planet, the whole city will have more validation. So we keep this framework directly for the cities. And on the side, we have more data sources, very quick review. Like we have more people and more devices are connected. So we have more data. They're increasing very fast. Some foreshadest of all the total amount of data accumulated by human. And besides the data, we also have devices that means computing capability and also some algorithm or models and so on. So with the same of data, machine and models, we can use all the information more impenetrable. So how about the data? We see the difference between the, we call it the new type of data because the data is not very specific. The traditional data is like some table or source statistics published by government. So the new type of data like cell phone, internet, open table. You have open table where we have similar company in China, like a trip advisor. People comments on the tourist, like the floating bike, like taxi. All of the data are related to the city problem. So they have higher frequency and final gravity and reach a property that's a little different between the new type of data and the traditional data. So the first of case is an article we published like when the four years ago. Different version because we talk about little different. We evaluate the old effect for each landfuse in China with social media data and land scan data. So I just go through this article. The story made you work on these articles about the way I mentioned the H2S emission. So we can calculate the range of the oath which maybe use merit. We use two different models like gashing, disparaging model and one to cut off more specific and disparaging model. And then my part is in that I use social media data which contains the location and text contain information to validate the range of the owners. So this is all the future of all the landfuse near into some landfuse. Also in the 55 landfuse in China, it's from my call, it's not from me. So you can see the different side of circle and because of the estimation, the H2S emission and they cover about a very easy part of the China's land territory. And the range, the calculation with the average radius of the range is about 800 meters. And this part is about social media. We extract the social media with a content of landfuse and us, like things of this sort of, to say where the social media occurred. Like the first one is Guangdong province, Guangdong Beijing Zhejiang province and the campaign, sorry, I cannot translate this picture. It's all about the topics and words that are talking about landfuse when they are using social media. So the red part is, you see the red circle, that's the result of Gaussian dispersion and the green part, most specific, from car proof, that's the result. I'm not going to that detail. And you see this map, every year or dot is a social media. So we collect all of the social media to say the distance between the landfuse and where people may post in a social media post. So it's very interesting map, which we can evaluate out of fact. The first one is Beijing, Guangdong, landfuse, and this is Nanjing, where if it's a mountain, people come to the mountain but just behind the mountain is a landfuse. We also use another data set, it's a land scale from NASA. It's a population data set for every two meters square. And so we could save out in fact, how many people are seen in China and what kind of the people has been in fact. So we have this issue, the land scale is more statistic than in the population, just a number. And the social media we see is more dynamic because you know where and when the post has appeared on the internet. So we will see the issue and the land carving about the ranch and the carving ranch and contain how many population of statistic and dynamic. So we see the effect, something interesting. And then I read another article to use that relationship, we want to make a tour for planners. That means if we know the effect, how could we plan the facility better because we are working for the planning institute. So the major work is we use different population which we have mentioned the land scale and the geotagging social media. And we have to evaluate two things, one is the overall pattern in a city about the landfuse and for every plan for each facility, each specific facility, how the, it's located good location or bad location. And we can compare different facilities within different cities in China because we use the same dataset. So this is Beijing under Shanghai, I take these two cities, for example, all the grid dots very, their entities, they are social media. So you'll see where the people active, they could post the social media because we collect the social media for a long time. And the black one is landfuse, landfuse. So this is the basic data. And we get a very simple model that's the inverse distance which is, that means if you have more people near the landfuse, the score is much lower. It's a very simple idea. And all of the data has been specially joined to a grid. And then here's some results. And I'm not going very deeply to that. I can share this paper, unfortunately it's in Chinese. Okay, we say two lines, one is the dash dot line and another is, and the off chart is Beijing and the below chart is Shanghai. So generally, if you near the one line is a medium, medium value, another is average. So if you near the landfuse, the media is always much higher than average. That means some landfuse, very few landfuse, they affect a lot of people. So makes the average much higher than the medium value. And we can have a ranking of the two cities. The ranking first one is Shanghai, Laogang and Fugh and the second one is Beijing, Fenghai. So we can compare different landfus from different cities. Okay, the second case, we do a proposal, it's not a project, it's a proposal for competition. Okay, we do a competition proposal for Wuhan on a similar lab. We propose a city greenhouse gas grid management platform. So the most work is about this four and point. And first, we have data for our GHG inventory for cities that come from my co-authors. So I'm not going to the details, and he is an expert. And we have some control of the quality of GHG estimation because he's working for a mixed trail of environment. So he gets some data from the online monitoring devices. So he can control the quality of the data simulation. And my part is about to make an interactive system. We can analyze all the results. We make a final gradient GHG grid in Wuhan, which means the spatial is much smaller, about one kilometer. And the time is every month, so it updates the data for every month. And the interface is much more modern and very easy to use, analyzing, visualizing. Different indicators in Wuhan define kind of the GHG, like from industry, or from residential, or from agriculture. And you can do some analyze by making a fork, that means you just click and get a result. We can see the hot code area of GHG admission. Some parts of the grid are very high and some parts are very cold, very low. We can see the pattern, generalized. I just click the mouse. Okay, so we'll have some other cases. It's not directly related to the waste that go into forest. We collect a lot of data in internet street view, which the Google map has the street view and the internet we have I do and the Google map that has street view. And we can analyze the content of the picture to see what's on the street. The building, the cars, all the skies, the ratio, how many ratio appear in the picture. And the dock is floating bike, which means no bike or all FO. Different bikes have traffic moving around the cities. And like, this is people are jumping, they're running, people are like running in the gym, and after they run, after they have the excitement, they share the tragedy also comes here. And then we got this tragedy trajectory to analyze where they have often runs. So this is the map. And then we can analyze to combine the trajectory and the street view, say where people are more like to have exercise outside. So there are some results, we do the two detector. Two detector is a method of which you can, detecting which factor are more important than to your attribute. The attribute is the amount of running or amount of biking. Like a precaution for the sky, people are like to have exercise outside and the open space dispense from the green kind of the factors. And even we also have some, we set up the monitoring by ourselves to get some air quality or we can count the people and count the vehicles as a very old district in Beijing. And then we do some tests for like a visualization or a very huge screen, and the Chinese people like this especially for people in government. And we have different examples of this. So we combine all of the information for their decision making to support that decision making to do some analysis on the talk. Okay, some final remarks. First, I see the triangle caused by Paul. They're a triangle system data and decision making, but I say this more like a loop. Loop, so it's observation, orientation, decision making, action. So run to make a close to loop, close the loop which you mentioned in your component. And for the data, I'll get perception, measurement, mining or analyzing in simulation. So it's another data loop. So all of our work is going to making more a close to loop for the city issues. And the first remark is a term like a CTS. It's a cyber-physical system which we have physical environment like cities. And we have some virtual cyber space to compare all of the data. And the city has some difficulty to use the data because the order is open. You cannot draw a line of the city. So it is important. I mean, cancer is not clear. So we have to make the connection to make more power of data for support of the physical improvement. Second is, we see the IT and DT is totally different. Okay, IT system is made by collecting and communicating without doing some analyze. That's for the data technology. We want to use data to support decision making to save the consequences behind the indicators. That's how I say it's data technology. And the last remark is data or say AI is not online. It has a lot of show cards. So we see these words like augmented intelligent. It's much better. AI is useful to help our world to have better thinking about the situation. Okay, thank you.