Great Toilet Simulator

[Lara Preece]

VRToilet Simulator is a free app for Android published in the Simulation list of apps, part of Games & Entertainment.

The company that develops VR Toilet Simulator is equbytes. The latest version released by its developer is 2.0.4. This app was rated by 1 users of our site and has an average rating of 4.0.

To install VR Toilet Simulator on your Android device, just click the green Continue To App button above to start the installation process. The app is listed on our website since 2017-01-02 and was downloaded 63 times. We have already checked if the download link is safe, however for your own protection we recommend that you scan the downloaded app with your antivirus. Your antivirus may detect the VR Toilet Simulator as malware as malware if the download link to de.locovr.vrtoiletsimulator is broken.

How to install VR Toilet Simulator on your Android device:Click on the Continue To App button on our website. This will redirect you to Google Play.Once the VR Toilet Simulator is shown in the Google Play listing of your Android device, you can start its download and installation. Tap on the Install button located below the search bar and to the right of the app icon.A pop-up window with the permissions required by VR Toilet Simulator will be shown. Click on Accept to continue the process.VR Toilet Simulator will be downloaded onto your device, displaying a progress. Once the download completes, the installation will start and you'll get a notification after the installation is finished.

LinkedIn and 3rd parties use essential and non-essential cookies to provide, secure, analyze and improve our Services, and to show you relevant ads (including professional and job ads) on and off LinkedIn. Learn more in our Cookie Policy.

TLDR: this post suggests a high-level architecture for a Coronavirus simulation model, defined in 10 steps. You can implement this model either as an exercise in data science or if you are in a position where you have real decisions to make regarding the response to the virus.

Saddening news about the Coronavirus spread fast, alongside videos of shoppers piling up toilet rolls and Italians singing in their balconies. This week there have also been some very good posts from data scientists, translating the mess of news items into new insights.

I'm a simulation geek, and spent years using different simulation approaches to support decision making and help reach optimal trade-offs between conflicting goals. This is a very timely topic, since every news source today is full of items about the tradeoffs involved in putting people in isolation. There are questions about whether, when and where to impose restrictions on people's activity outside home, and it's generally agreed that we'd like to balance two contradictory objectives. One objective is to minimse the spread of the virus. The other objective is to minimise the economic, social, mental and other impacts of restricting people's movement and putting essential activities on hold.

There is a very high degree of uncertainty in the current global state, since the virus follows an unknown behaviour. Fragmented and noisy bits of data flow in all the time from different places where different policy measures were (or weren't) taken. This is actually the classical situation where you'd use simulation. You build a model that estimates certain outcomes as a function of many unknown parameters, with dozens of assumptions about cause and effect relationships. The model by itself is a weak tool, and won't directly tell you what you need to do, but...:

1) As data flows in, you can keep refining your model in real time, and learn something from this. Every day you have a slightly better model. So you can re-run the model every day and re-assess your strategy continuously.

2) The purpose of the model isn't to simply tell you what to do. The purpose of the model is to support your scenario analysis so that, after examining different possible situations, you know which strategies increase the probability of an optimal outcome, together with evidence from other sources.

5) The model can focus on non-obvious strategies rather than the black-or-white ones. It's not just "business as usual" versus "complete siege"; maybe the conflict between the objectives is minimised if we close down public buildings for just 8 days? Maybe the spread of the virus can halve just by targeting rail travellers that leave home before 10am and go somewhere for longer than an hour?

6) The model can focus on those specific things that YOU need to optimise. Not just whether you close all airports in your country, but also what to do about a specific workspace or whether to attend a wedding in a remote village.

I wish I could take a few days off, leave everything else and build this model. I won't be able to do it, but in case you can, below is a recipe for you. This is a proposed high-level model architecture which you can implement. If you are self-isolating at home with a bit of free time, or an acedemic looking for your next prediction piece, or a data scientist between jobs, or a consulting group with a research fund - please grab my recipe and go build a model based on this suggested architecture.

Oh, before getting into the recipe, I'd just note here that I'm clearly not a public health expert. There are experts out there that specialise in the statistics of diseaese spread, who already have much better models than anything I'd come up with. I've seen some interesting simulation work by the WHO, for example. Still - read the news and you'll realise that decision makers don't necessarily follow what experts say, and the experts themselves have a wide range of views, so the recipe below is my humble contribution for you to use in whichever way it helps.

My recipe is for scoping your model in 10 steps. It's a relatively simple model - you can write it as one programme in Python or any other programming language. It would take a few days to build but only seconds to run, so once it's built, you can look at many scenarios immediately. I've no idea what conclusions it will help you reach, but if it doesn't help, it probably won't harm.

1) Geographical scope. Define an area that can make its own policy decisions and has its own distinct culture. A single country would make sense as a start. Ideally you should break it down into smaller areas, so that you can add population size per area, distance between areas, and level of interaction between areas. I would add an international dimension to the model structure too, but a bit later, so I suggest that you populate only one country as a start, but use a data schema that can allow multiple countries in later versions of the model.

2) Agents. I'd use micro-simulation, i.e. simulation where the agent is an "atomic" entity that can make its own decisions. The natural choice would be that every person in your geographic area is a separate agent. Each agent has decisions to make (e.g. do I go to work? do I go to the shop? do I sing cheesy operas in the balcony and put it on YouTube? Ok, maybe not the last one). Each individual would have a state, e.g. "healthy", "carrying the virus", "ill", "critically ill", "dead". If you define individuals as your agents, each simulation step will need to re-calculate the state of each agent (see more about this later). Alternatively, you can use a model where the agent is an aggregated group of people, such as a household, a ward, a city. If you follow this approach, it would be harder to simulate the decisions made by individuals, since your agents aren't atomic decision-making entities, but you can get away with it by simulating the distribution of individual decisions across this aggregated group of people.

3) Segmentation. You'll have to divide your population into segments so that people in different segments can follow different typical behaviours. You'll need to consider data availability, so that you have engouh data for each segment, although this recipe assumes that you pick most of your data from what's published in the media. Age seems the most critical segmentation criterion, since the virus affects different age groups in very different ways. Underlying health condition would also be an important criterion, but harder to find enough data about. I would actually add the underlying health as a segmentation criteria even if I had no data at all and I had to guess different proportions, based on general knowledge; this would allow running the model under different assumptions of how these segments split, and improving the assumptions as part of the model training stage. Another important set of segmentation criteria would be the mobility habits of your agents: how many travel to work daily by train? how many travel by car? or are home-based? This affects their likelihood to spread the virus. Finally, another segmentation criterion is the degree of daily interaction with others that comes with their occupation. A simple model would only have two or three levels of this criterion, low / medium / high. A more advanced model would incorporate the interaction with others in more quantitative terms - how many people they interact with, at what frequency, at what level of physical proximity, and so on.

4) Objectives. To run the model itself you don't need objectives - you just examine how the outputs evolve under various policies and assumptions. But you might want to define objectives anyway, so that they guide the way you change your policies across different scenarios, i.e. between model runs. We already said that the main global objectives are to minimise the spread of the disease and, at the same time, to minimise the socio-economic cost of measures that restrict public activity. If you want, you can break these objectives down into more specific ones, because there will be many types of impacts: finance; crime; environment; well-being; education. You can also add local objectives wherever you want to use the model to help with your own specific dilemmas. For example, you may decide that the health of medical stuff is of the greatest importance, since other people's health depends on their services, so you can add the health profile across this specific segment as a separate objective. Similarly, you might work for a business, whose viability depends on the availability of your field staff, those with much public interaction; you can define a separate objective for this.

3a8082e126