Big data gives us nine "lessons"

Yesterday and today I attended the Ohio State University's Big Data Future symposium. Ohio is working with IBM to create a large data center on the ground, becoming an important academic force in the field. The symposium brought together experts from across the country and an expert from the UK to conduct a fruitful discussion on current achievements, trends and topics in the field. It is a pity that I did not attend this afternoon's exhibition, but you can learn more about it on the website of the Conference and the forthcoming publication of the journal Law and Policy of the Information Society (ajournaloflawandpolicyfortheinformationsociety).

Big data and open data are not the same thing, but they have a close connection (as I said in the topic "How far will the big data of the future be open?"). ). Some of the big data we're looking at are also related to open data. In this vein, I came up with the nine things I learned about big data on my way to Columbus. On the official website of the seminar, I can see the names of people in bold.

Prepare for a drastic response to large data. Many speakers have referred to the topic of "Big Data hype", thinking that big data is being discussed so enthusiastically that we can now go into a reactive loop. Mikenelson the "rubbish data" he saw in public, and even suggested that we should rename the big numbers, which could have a "big Brother" – like many people. He suggested renaming it: Bffmudd, an abbreviation for large (big), fat (fat), fast (fast), disorderly (messy), unstructured (unstructured), distributed data (Distributeddata).

Aware of "big data hubris". Several speakers quoted a new report showing that the "Google Flu trend" – one of the first big examples of big data predictive value – proved to be very inaccurate. Clearly, Google might be smart to tweak its algorithm in a wrong way. Whatever the mistake, it's a lesson that if you don't look at the broader picture and just try to find the truth by crushing the data, you usually don't get the desired result.

Data is not a substitute for judgment. Data, especially large data, are tools that can help people make decisions, but they do not act as substitutes. Rayharishankar said: "Data plus analysis is information, information plus context can provide insight, insight will be able to guide the right action, the right action will bring about the result of the value of Ascension."

Correlation cannot be stronger than theory. Some of the big data advocates argue that big data almost makes theory redundant: they say that with enough data, we can find many important and useful patterns and trends, even if there is no theory to explain why. Indeed, simple correlations can, to some extent, drive accurate predictions. But even with the ability to predict, it doesn't mean you can really understand how the system you're studying is running. Eytanadar suggests that we look at all the relevant efforts of large data ranges from predictability to interpretation, and focus more on understanding what we see, rather than just focusing on patterns that can predict the future.

Big data is-risk-tracking a "mobile" society. On a global scale, mobile devices have become the preferred online connection tool for humans. Farnamjahanian points out that by 2015 the number of mobile devices worldwide will be twice times that of the population, all devices can send location information and other data to companies that can collect the data. This will be one of the major sources of future social data. But Katecrawford? points out the privacy risks here: Because of the uniqueness of human mobility patterns, you can identify a person by using only 3-4 of the data points generated by the mobile phone.

Big data can help-or damage the city's democratic institutions. As Harveymiller said, the ability to track city activities through mobile data, Remote Environment sensors, laser-generated aerial maps, and more tools can give us a super coordinated city with a higher metabolic function. (Unfortunately, I had to leave before Michaelbatty's keynote speech on city analysis, but he offered a speech on his personal website) but Katecrawford again warned. If we are not careful, urban data collection will be asymmetric to help the rich and hurt the poor. Boston's Streetbumpapp app, for example, collects data from potholes by tracking the swing status of smartphones, using volunteer data to reflect the bumpy bumps of a road. But most smartphone owners belong to affluent people who initially monitor and repair potholes in richer areas – a problem streetbump is currently working on to fix. On the opposite side, "predictive monitoring" is being used to enforce police control in areas where high crime rates are expected, leading to discriminatory enforcement.

Privacy still matters. Forget about reports that the public, especially young people, have abandoned their privacy. We still care about privacy, but we just don't know what to do. Here are two things to consider: we want to know what information the government agencies or data-tracking companies have gathered about us, and if we don't like it, we want them to stop collecting. How to solve these considerations is not clear. Some speakers at the meeting suggested a simple solution: to make governments and companies more transparent about the data they were collecting, something that some people call "mutual recognition". But a long-term transparency advocate, garybass, says the proposed solution "is not the real world." For the past 30 years, I've struggled to get data available, and governments and companies are desperately trying to make data inaccessible ... This is a protracted struggle. The risk here, as others say, is that we may enhance the asymmetry of power between the data collectors and the collectors.

Large data should show the beauty of the data. Rapid advances in data visualization are creating stunning results. For example, take a look at this "experience bike crowd" video, gradually analyze the London bike traffic data, show the Ohio supercomputer Center the clearest model and partially completed visualization. Such data visualization is not just about aesthetics, it's about understanding. Angelashen-hsieh, a data visualization expert at IBM, talked about the need to make data "more fit for human consumption" and to focus on "the last 18 inches" from the computer screen to the human brain's messaging journey.

Large data will (most likely) produce great value. Aside from all warnings, there are many social and economic values in large data that can be exploited. McKinsey, a landmark big Data report a few years ago, predicts it will pry into the economic value of trillions of of billions of dollars. Angelabyers, a co-author of the study, said today that it may still take another 5-10 years to produce this value, in part because we still face an important skill gap: The number of data available and the gap between the number of people who can make use of the data. But economic values are emerging and in some surprising way. Johanbollen and his team used the big data sentiment analysis on Twitter to predict the stock market: They calculated Twitter's "calm" mood to predict the closing point of the Dow Jones three days later.