Welcome to this special session of the MOOC in platform thinking. Today, I'm very glad to have here as a guest Alessandro Tansini and the very first question I'd like to ask you Alessandro is if you can tell us where you work and what's your background. Hello, it's a pleasure to be invited here. I'm Alessandro Tansini, energy engineer with a background on power production and automotive systems and for the past five years I've worked at the European Commission Joint Research Center. You know, this kind of question usually is kind of easy in this kind of sessions because people that are being interviewed here are working for digital platforms, the topic of this course. Well, you have a completely different background and you work for an organization that doesn't really look like a digital platform and the reason why you are here is for something that happened a couple of weeks ago. Actually, the two of us know each other, we were talking at the phone and something kind of strange happened. He asked me what kind of car I was driving and when it was certified, and I didn't know I had to search for it and at that point he started asking strange questions at least to me regarding my willingness to have this object in my car for a bit of time. Alessandro, can you please tell to the people listening to us why were you interested in those features of my new car? Sure. As said, I don't work with digital platforms, I work in a research center where we deal with sustainable transport and we are interested in monitoring the real world fuel efficiency and energy consumption of passenger cars. Not only passenger cars, also heavy duty vehicles sometimes and commercial vehicles, but this specific device is developed to be used on passenger cars and this specific device monitors the private use of cars and why do we need that? Every vehicle, when certified, is type approved with a nominal value of fuel consumption, so when you have to buy a car, you check online or on the material provided by the manufacturer and you are told that your vehicle will consume a certain amount of fuel in mixed condition or urban condition, highway conditions and you trust these values. Just to bring it to a very simple level because it's kind of out of topic for this setting. You're talking about the number of kilometers that you can do with a liter of fuel for example. So the values that you see when you buy the car that might be kind of similar or not when you actually drive it, that's what you meant right? Exactly. To bring an example, when you check for the details of a car, you might read somewhere this car consumes five liters per 100 kilometer of distance driven, so this is the nominal value, but this might differ from reality because this data, this specific value, is derived under certain conditions obtained in the laboratory and, in real world, you might have different conditions like climatic conditions or use conditions that might cause the difference, so with this device we want to monitor the real world fuel efficiency of vehicles and this is important for us to monitor the efficiency or better effectiveness of a certification procedure for vehicles. We can monitor if there is a gap and how big it is between the nominal value and the real world value, so if you consume more than the nominal value, we can understand with this device and hopefully we can also understand what's the factor that causes that deviation. The first thing that you asked me when we started talking about this thing was when my car was certified. Why does this matter? So, this specific feature of my car was relevant? It matters because I needed to check on the availability of a feature that new cars have. New cars certified with the latest emission standard euro 6 AP (but depending on the country there might be different acronyms), with this latest emission standard vehicles need to store in the control units some pieces of information, like the total distance driven, the total amount of fuel consumed in the lifetime and if it is a plug-in hybrid electric vehicle there is additional information, for example the amount of electric energy that was charged into the battery. So because of this feature data is available in a standardized way in country units of the vehicle and these data can be extracted from the vehicle. So, I needed to know whether your car was compatible with this new feature and this is something that you confirmed me with the date of production and also the acronym for the emission standards and then I knew that I could use your car for the project. Okay, so you are telling me that my car is generating data on what the car actually does, it's storing them and it's doing it all by itself by default, that's something that is happening by law and it's happening thanks to a technology that is within the car. Exactly, the technology is nothing new but every manufacturer was doing this in a custom way, let's say, but as of 2021 with the latest emission standard, this is standardized, so every manufacturer will have to implement this feature following specific rules and complying with accuracy requirements, so I can plug this to your vehicle and extract the information, I can plug it to the future vehicle I will buy and extract information in the same way. So, that's what's new and the acronym for this new feature is OBFCM, onboard fuel consumption monitoring, and this is what we take advantage of for our project. Okay, so you are telling me that all the car manufacturers are actually doing this in the same way, so all these data will be comparable, can actually be used together and so on and so forth. Is this something happening only in Europe or is this happening also in other territories? To my knowledge so far only in Europe. So it's a European regulation, so far. Yes, the regulation is surely European. This specific technology or, let's say the format of the signals might be available in other countries. I don't know if in other countries this is mandatory, it has to be monitored like we're going to do in Europe and so on, but definitely there's a European regulation forcing manufacturers to implement this feature and as a research center of the European Commission we will be looking into this data. Okay, so you are telling me that starting today I will have this in my car. Can you tell me something more about what it is, what kind of data are actually available, gathered through this thing and how it works? Sure! Although very small, it is packed with a lot of hardware, it's open source hardware with the open source software that we developed to our own requirements and needs and desires. Is it something developed by your center or it's developed for centers like yours? It's open source hardware that you can buy online and there is a person who gathered all the hardware and packed it together in this enclosure. We bought this enclosure with hardware and we also downloaded the open source software made available by the developer but then we are developing the software to our needs, let's say, so we have implemented the new feature that I explained before, OBFCM data collection, so that this device, which initially was not able to gather these data, is now able to gather these data and this device sends data to the internet, so I will explain a bit more how this is achieved. This is an arduino compatible device. It has an arduino compatible board with another microprocessor that connects to the onboard diagnostic port of vehicles. Which is the one that people use when they are fixing your car, right? Exactly, they are used at the mechanics or when the they call undergoes service and maintenance. So we take advantage of the same port for communicating with the vehicle and the microprocessor is developed in such a way that sends the correct request to get OBFCM data; the data is returned from the control unit, the OBFCM data; then this device also has a cellular module and a wi-fi module so that it has connectivity and it can connect to our server so that we get the data that is being read instantaneously, so real time. It also has a GPS sensor, so we get the positioning and other sensors from the device, like temperature, voltage (this is to track how the the device is operating). So, basically this is reading data from my car and adding somehow other data to the ones you'll be receiving, so you know where I am thanks GPS not because of my car but because of this sensor, right? Yes. So positioning comes from the GPS module that is embedded in this device but from OBD on-board diagnostics we get both OBFCM data, so the onboard for consumption monitoring data that I introduced before but also other instantaneous signals that the car makes available for other reasons, which is vehicle speed, maybe ambient temperature, accelerator pedal pressure (so I know how much you're stepping on the accelerator pedal), might be engine load, might be information about how the vehicle is running, for example the rotational speed of the engine. So we get quite a big amount of data which we use to characterize the driving behavior of citizens not because we want to track citizens but because we are interested to understand in the fleet at country level how vehicles are used. So for example the average driven distance in one day or the average velocity. It's very interesting and it's generating another question, which is, how many people are you expecting to convince to use this kind of device? Not obviously the exact number but an order of magnitude for this project. We are still in a very preliminary phase. This is the first deployment of a technology like that and a device like that. We have about 30 devices which we are distributing to people that somehow we know because at this point it's a bit hard to find people that are willing to have a device in their car without knowing who exactly is getting the data. So we try to reach out to people that somehow know what we are doing, that we are not doing it because we want to monitor their private life, but we do it because of our needs at work. So far, we have distributed about 10 devices more or less in different countries in Europe. Hopefully, we are going to distribute another 10 very soon but when everything is in place, then such a system, so a device that sits in the car and connects remotely to a centralized server to communicate this data, when this is in place we want thousands of vehicles to be covered by this project, obviously in a more structured way with encrypted data, so that privacy is ensured. That's interesting, so this is kind of the startup phase of this project, but you're planning to have it definitely wider. When you presented this to me, we were at the phone so I couldn't really see this object but actually I kind of remember it, even though I've never seen this one, because when we start doing research on platforms it was 2014. There was a case, that is called Automatic, that is to me at least pretty similar but white and from the United States that actually had this connector for cars and was promising to people something slightly different. That's a B2C, so that's a consumer product that was aiming to tell people something more about their driving style, just as example I remember from the advertisement of this product: if you push the pedal in different ways, you can save a lot of fuel and if you have someone that constantly look at how you're actually doing it, it can give you practical advice on how you can help the planet and save money using less fuel. I never bought it but I was very interested in it from a research perspective, because this was very similar to what we were doing with our research, so these companies that are gathering a huge amount of data and this data can be valuable for someone else. When I told you this story, when we were chatting at the phone, you told me something pretty interesting so I never knew if someone else, just some companies, were actually on the other side of this possible platform which can be Automatic and what's your take on this? What do you know about that kind of company? It's not Automatic in the specific but those companies that do this kind of services? For us, they are a gold mine because basically we are asked to support EU policies by having an understanding of the real world, but we do not have eyes everywhere and if anybody else was able to put their eyes around then we are very willing on buying that data that let us better understand again mobility needs and what is happening out there in the world, because we have certification that defines nominal values for things but then you have reality and you need to understand how is reality, if this differs and how much from the nominal values for things. Our project is on cars but there are other examples to that. For example, we are very much interested into monitoring the gap between nominal fuel consumption values from cars and right now that we have electric vehicles also energy consumption of electric vehicles and monitor the difference with what you achieve in real world. So this is the gap and data sets, like the data set you mentioned from this company, is something that lets us understand in one instance what is the average fuel consumption value of a circulating fleet of vehicles in Europe in different countries, which is something that we also want to understand because if we carry out our own test activities maybe we will reach out to people that are close to our area, so this is a biased sample or a subset of the population you would like to track. So, if somebody was able to reach out to many people and collect data in a diversified way from many sources, this is something that is very valuable to us as a research group that needs understanding of what's happening in reality. That's interesting, there is a customer and actually I've been knowing one and I didn't know and you're telling me a couple of things more with this, which is there are other companies doing something similar to what Automatic is doing? Yes. And are they offering kind of different services to the user ? I have different user experiences in comparison to the one I mentioned ,do they do different things for for the final person? Yes, I can speak about the ones that we use, that we took advantage of and these specific services were about fear consumption monitoring and reporting, so some sort of website or smartphone application where users would insert the driven distance. The fuel they would use also may be something about speed and one of these services had something like you say. So one of these devices connected to the vehicle and constantly sending information over to their server, so they are able to derive every quantity they would like to derive or simply sell the whole data set to other research centers that need data for their own purposes. So you are telling me something more. There are companies that are actually doing something pretty similar, so there is a device. They are automatically gathering a huge amount of data for certain purposes to the customer, they gathered them and you actually buy those data sets and there are other services in which is the user that is somehow creating those data. You said there are websites where you put for example in the card the amount of kilometers, so obviously those data are less rich in comparison to the others but you even need that kind of data for your research. Yes. And on top of being less rich as a data set, this is also subject to human error when inserting this information, so this is a bit trickier for us when we have to get rid of outliers and we have to normalize the data before using them for our needs. It's very interesting to see how a research institution that actually help in defining regulations needs data to take decisions to understand things and to get these data is so difficult that actually you start from services that can gather them manually and I can only imagine the amount of opportunities that this is opening. What you were saying brings us to a different aspect of this conversation. You were mentioning about companies, various kinds of services that do gather data to offer something to an end user, to a customer and then you are on the other side talking about platforms gathering and using these data for research purposes. You work for the European Commission so I'm absolutely sure about the answer you are going to give to this answer, but I want to make it anyhow since it's a very important topic in this kind of discussion. Do you have any access, any view on who generated those data. What kind of data you actually see when you buy these data sets? No, we don't have that type of information, we don't receive personal data from the people who joined these activities. We get aggregated data that are also anonymized and we are barely allowed to understand what type of vehicle was used, so we maybe get only the information whether it is a diesel engine, gasoline engine vehicle and few other characteristics that are relevant to the vehicles, not to the user. Okay, and is this generating any kind of issues for what you are doing, so do you care anything about who is the person that generated those data? No, as we said before we care about data collection from the whole community, so we collect data and we treat cases as a whole. We want the global picture so we don't care on the one by one case. So the more data you have, the more valuable they are, but you don't care anything about who is behind the single data point. And this leads us to the last part of our conversation. You know, we've got a personal relationship so probably this conversation was kind of strange, but when I said: okay, you can have my data in exchange of this interview, you started asking things like: but if you wish, I cannot track your position; if you wish, I can switch off some part of the data that I'll be gathering, and so on and so forth. And you know, recording these kind of courses and talking about these kind of topics it would be pretty strange at least to me saying: yes please, keep that away, I don't want to share them especially with you. But what's the typical reaction to this kind of questions? How is it finding people that do participate in this kind of project? As long as we reach out to the eco nerds, how we call them, it's very easy to get fear consumption data. So, it's people that typically take care of how much fuel they are consuming, so they are a hundred percent open to share these data and also they can track in a better way how much fuel they are consuming, but then we also say that this device is not only recording fuel but also recording how you drive, so if you comply with the speed limits on public roads or where are you going when are you taking your car, is it day, is it night. So some people react badly when we say for example that we're going to record GPS positioning, because this really enables us to understand how the vehicle is used and what a person is doing in its private time, but as said before, we treat data as a whole and anyway in order to reach out to more people. To get more volunteers for our project we provide different options for joining us, so if they are completely open, we ask to get the whole set of data because that's very useful to us. We put more context to our data, so in case we have anything to analyze deeper, we have data on the side that might help us to better understand something about fear consumption mostly but if people do not agree on sharing positioning then we can get rid of that and only monitor the driving signal, so for example speed or accelerator pedal engine, RPM ambient temperature. The very smallest set of data we offer is only that from OBFCM, which are something like counters that they are just accumulating, but we lose the second by second picture of driving on public roads, so this is only to reach out to more people because we understand that some people might not feel comfortable with sharing that amount of data so in order to still collect the data we really need, which is the average fuel efficiency and energy consumption from vehicle, we are also okay with receiving a smaller data set. You are saying two amazing things for people who talk about platform: the first thing was related to the people that more easily accept to share these data and you are calling them eco-nerds, so people that are interested in the kind of work you are doing probably it's not enough to have them on board, because you need a wider picture. And we cannot only have eco-nerds, otherwise that would be a biased sentence again. We have to reach out to some eco-nerds but also to non eco-nerds. But there might be somehow a way to connect what you are doing with the people that actually create data, so this is kind of interesting and the second thing was related to the fact that you can actually leave freedom to the end user in your experiment, in your project to actually decide which data to put there or not and this is something that is not very common from a user perspective or, better, it is common but it is usually pretty hidden behind many features of the service, but it's very interesting to see that in a startup project basically it's something that you put at first, basically. You were mentioning something that is in my opinion potentially very very interesting. You said: the more data you allow us to get, the more we can actually understand about the consumption of the car and you said: if I have the GPS system, I know where you are basically. So I was wondering, can you use this set of data to connect what is happening within the car with what is happening outside the car with other databases that are not related to these objects, but can be available online from other providers or things like that. Definitely this is something that we do and the reason why we ask for GPS positioning, for example, is that we are interested or we need to know if there is any road slope to account for, so your fuel consumption increases when you're driving uphill and decreases when you're driving downhill, so from latitude and longitude we could query services that return altitude profile of the trip that the user is driving but also we can connect this set of data to climatological data so that we can also derive other quantities that we are not monitoring. We are monitoring for example average ambient temperature, but maybe also humidity could affect fuel consumption. You have to use your air conditioning system a bit more, maybe if the humidity is high; or solar radiation (this maybe will affect in a way that you have to use more the air conditioning again). It means that not only you know that in my car there was a certain amount of degrees but you know even the weather outside, the degrees outside and so on and so forth, if you have the GPS because you can use that as a linking point, right? Yes and we can do this already and something we might be doing in the future that we don't do right now is getting data on a traffic, for example this could also be relevant to fuel consumption analysis. Okay, so this was probably the most peculiar case we have in the entire course, something that doesn't look like a platform and indeed, so far it isn't because you are looking at the world of platforms from another site, but it is extremely interesting to see how research can be powered by data that are gathered in this case through a prototype, an initial project, but that potentially can scale up and leverage the data of thousands of people even for the greater good of society and for the environment in this case. Thank you very much Alessandro for being here with us and for sharing your knowledge, your experience and your project with us. Thank you.