Big data cases 21-30: Opower contrast incentives, Takadu's digital watering

21 Takadu's digital watering

Water has always been a bad thing to manage: the water company found that there was a problem with a water pressure gauge and it might take a long time to check a number of water pipes sharing a water pressure gauge. When it was found, a lot of water was already wasted. An Israeli water system warning service company called Takadu solved the problem. Takadu collected the water pressure gauges, water consumption and weather data of underground water pipes, and sent them back to Takadu's computer through Amazon's cloud service for algorithm analysis. If there is a water pipe leaking or seeping in the underground tap water pipe somewhere in the city. As well as abnormal conditions such as insufficient water pressure, it will take about 10 minutes to complete the analysis and generate a report, which will be sent back to the maintenance department of the tap water pipe. In the report, in addition to providing the type of abnormal conditions and the damage of the water pipe - how many cubic meters of water leaked per second, it is relatively accurate to indicate where the problem water pipe is. To detect the "waterway" per kilometer, Takadu's monthly fee is $10,000.

22 Lily Net's Marriage Match

The cash income of the e-commerce industry is derived from data, and the business model of marriage and love websites is rooted in the study of data. For example, as a marriage and love website, Lily Network not only needs to do some research reports, but also analyzes the age, geography, education, economic income and other data of registered users. Even if it is a small avatar photo of each registered user, there is a lot of excavation behind it. the value of. Li Qi, research and planning department of Lily Net, once analyzed the avatar information of the massive registered users on Lily. It was found that the photos of popular avatars are not only related to the appearance of the photo owner, but also the expressions, face proportions and clarity of the characters on the photos. To a large extent, it determines the popularity of the photo owner. For example, for female members, a smiling expression, a direct look at the front and a faint makeup can increase the probability of their popularity, while those with a face proportion of 1/2 of the photo, wearing formal, and looking directly at the eyes without a superfluous pose are More likely to become the darling of the marriage and love website.

23 Prismatic's press coat

Prismatic is a personalized news application with only 4 founding employees, relying on Amazon's cloud computing platform to realize the lean startup of big data with the data of Internet data crawler and social network development platform. Instead of providing a unified, well-designed news feed or referral interface, Prismatic does one-on-one data analysis and recommendation for users based on analyzing their Facebook or Twitter profiles. From the perspective of profit model, Prismatic does not rely on advertising fees to survive, nor is it a traditional news media, but an e-commerce company draped in a news application. In order to personally recommend news to the user, the user's personal information is analyzed for data, and the recommended product is launched in a targeted manner, thereby making a profit from e-commerce.

24 Opower contrast incentives

Human beings have the nature of comparison with the same kind. For example, a government agency collects data from multiple groups of employees working in different locations in different locations. Just making this information public makes the backward employees improve their performance. In the energy industry, Opower uses data comparisons to improve the energy efficiency of consumer electricity and has achieved significant success. As a SaaS innovation company, Opower works with a number of power companies to analyze the cost of household electricity in the United States and compare it to the surrounding neighbors. The serviced families receive a comparative report every month. From the level of household electricity in the entire region or similar families in the United States to encourage energy savings. Opower's services cover millions of households in the United States and are expected to save $500 million annually in consumer electricity consumption in the United States. Opower reports envelopes that look like bills, and they use behavioral techniques to gently persuade utility customers to reduce consumption. Opower has launched its big data platform Opower4. By analyzing various smart meters and power usage, utilities such as power companies have become Opower's profit sources. For the average user, Opower is completely free.

25 Chango and Uniqlick click consumption

Using new data technologies, such as Chango in the US and Uniqlick in China, are exploring new business models in the digital advertising industry – real-time bidding for digital advertising. By understanding the behavior of Internet users searching, browsing, etc. on the web, these companies can provide advertisers with a group of users most likely to be interested in their products for precise marketing; a longer-term trend is to place ads at the most likely The user group purchased. This approach allows advertisers to achieve higher conversion rates, and for websites that advertise, it also increases the value of the ad slot.

26 Analysis of the behavior of public interest

Interest is the first social media data management platform in China. At present, the main social open platform in China is still relatively conservative in terms of the openness of user data. As a third-party data analysis company, the user data that can be obtained is still very limited. These user data are subject to user license. The interest is to filter the user data through the related data analysis principle such as operational statistics, and finally complete the description of the individual characteristics of a user's behavior and actions. These descriptions can help brand marketers understand consumer spending habits and needs; they can also help corporate leaders enhance their understanding of their employees. In addition to describing individual and group behavioral characteristics, these data analysis results can also be used to predict behavioral behaviors of user groups, providing marketers with some forward-looking market analysis. The results of the analysis of interest data can only be accurate to the group and cannot reach the individual. In addition to having a certain reference price in the field of marketing, most of the user data research of this type is mainly used to cooperate with some small research. In addition, these data can also achieve credit ratings for users and even corporate institutions, and also have a certain degree of use in the financial sector.

27 hauling the net tomorrow's guess

The drag-and-drop network of the shopping guide e-commerce made the application of "what to wear tomorrow." In this application, many fashion circle authorities convey fashion matching and style items, which can be arbitrarily scored by users. According to the user's preference, the net can guess what they want to wear tomorrow, and then she is in hundreds of thousands of online shopping. Recommend items in fashion, and achieve direct purchase orders. After getting the customer data, the background analysis is also a magical power.

The dragging net has added more variables to assess its recommendation mode. For example, some consumers will attend a party tomorrow, do not know what style to wear, nor look at the weather forecast, I hope that the shopping guide website can help her combine these scenes with their own information, giving a complete set of solutions. So date, region, occasion, style, these have become the variables of the dressing and matching solution, and are presented to the user through continuous combination. According to the data of the dragged net, the user sees a better quality match and has scene guide. , click to the last page to complete the purchase conversion rate will be 40% higher than the single product recommendation.

28 SeeChange's Gene Health

The ability to serialize human genetic files is now available, allowing doctors and scientists to predict the patient's vulnerability to certain diseases and other adverse conditions, reducing the time and expense of the treatment process. SeeChange, based in San Francisco, created a new health insurance model. By analyzing customers' personal health records, medical reimbursement records, and pharmacy data, the company judges the customer's susceptibility to chronic conditions and determines whether the customer is likely to profit from some customized rehabilitation packages. SeeChange also designs health plans and sets up incentives to encourage customers to take the initiative to complete health initiatives, all monitored through their data analysis engine.

29 Image diagnosis of Given Imaging

Israel's Given Imaging has invented a capsule with a built-in camera. After taking the patient, the capsule can take pictures of the digestive tract at a frequency of about 14 photos per second, and at the same time return the external image receiver. The patient's symptoms pass through the matching. The software is entered into the database and the capsule camera will be excreted from the body through the body within 4 to 6 hours. Generally speaking, doctors are judged by their own personal experience, and it is inevitable that they will not be able to delay or even delay the treatment of some suspected shadows. Now through the Given Imaging database, when a doctor finds a suspicious tumor, double-clicking on the current image, similar images taken by other doctors in the past and their diagnosis will be extracted. It can be said that the problem of a patient is no longer seen by a doctor, but thousands of doctors give advice at the same time and are supported by images from a large number of other patients. This comparison of data not only improves the efficiency of the doctor's diagnosis, but also improves the accuracy.

30 Entelo's "Before Headhunter"

Real technical talent is always a hot commodity for major companies. Don't wait for them to send you resumes, because they may have been snatched away by other companies before they have a chance to write a resume. Entelo can recommend high-level technical talents for entrepreneurs who are just motivated to change jobs, so that they can get started.

There are currently 300 million resumes in Entelo's database. And how to judge the tendency of senior talents to change jobs, Entelo has a set of patent-pending algorithms. This algorithm has more than 70 indicators for determining the tendency to change jobs. A company's share price decline, high-level exchanges, just being acquired by another big company, these will be seen by Entelo as a possible factor that causes the company to change jobs. Entelo will immediately push the information of the senior talents in the company to the entrepreneurs who subscribed to their services. The resumes received by entrepreneurs are not the same as the general resumes. Entelo has captured the information of these talents on major social networks. In this way, the entrepreneurs can understand what kind of code the person has submitted, what questions they have answered online, and what kind of information is published on Twitter. In short, these entrepreneurs who are ready to “dig the corner” can see a living target person standing in front of them.