MIT researchers have created a new analysis of the Urban Network (AN), a toolbox that allows designers and urban planners to describe the spatial patterns of cities, using mathematical methods of network analysis. These tools can support urban design, better informed and more resilient, and planning in the context of rapid urbanization. "The measures are useful predictors backbone for a number of interesting phenomena Read more [...]
Twitter has launched a new analytical tool aimed at measuring the influence of this channel in web traffic, and effectiveness of integrating Twitter into the sites. According to indicate the product will: Understand what content is spreading and sharing via Twitter network know the volume of traffic to your site sends twitter measure the effect of integrating the "Tweet Button" More info on the developer blog twitter: https://dev.twitter.com/blog/introducing-twitter-web-analytics Read more [...]
It is not yet well known, but Google has a predictive API, which implements algorithms "pattern-matching" and machine learning: Work against data sets that we provide and has libraries for sigientes languages: Java PHP Python Ruby. NET JavaScript oacurl Proposed use cURL R (at the moment I reserve judgment) Prediction of pre-emption of products and services based on histyórico Purchase Analysis Automatic document classification of posts (sentiment analysis) Guess how much a user might spend Read more [...]
Within the increasingly broad overview of Web Analytics tools, after an evaluation process of several months I recommend Piwik. Piwik is an open source alternative to Google Analytics, with the advantage of real-time web analytics and full access to data views in MySql DB, and integrates easily on the web and other applications that need this information. Besides the design flexibility of reporting, is a powerful analytical API: Piwik Analytics API. It is a tool from which to analyze multiple sites and has clients for Android. If you are heavy Read more [...]
In a previous post pointed application areas of data mining techniques with strong growth potential within the scope of "health", in sectors such as insurance, pharmaceutical and public administration. The following is one of many examples that will see light in the coming months. In this case, the use of scoring algorithms to predict success / failure of new compounds for drugs. You can view the article in MITnews. Although in its infancy from the industrial point of view, the use of these techniques can accelerate the drug design "smart" algorithms using not Read more [...]
The "Technographic Profile" Forrester classifies the participants in the online world in 6 exciting levels, depending on the level of interaction. The next tool available in Groundswell shows the differences by country, sex and age: The following presentation explains well the detail: The goal is to run a online CRM strategy in terms of role or level of interaction you have a person in the online environment. The idea seems very good ... from the theoretical point of view ... but is a little segmentation model operated from a marketing standpoint and also suffers operational Read more [...]
"The Internet Activity Index (IAI) Provides a unique way of looking at consumer engagement online, dividing Internet usage Into five distinct activities: Content, Communications, Commerce, Community and Search, defined as follows" ...http://www. online-publishers.org/page.php/prmID/421
A few days ago commented on the possibilities of application of advanced analysis in the field of eHealth enhanced by the development of information technologies. One of the first steps to obtain results in this area is to collect the patient's clinical information, to then share with doctors, clinics and hospitals in case of need. Well, I've found that Google has already taken a first step in this direction, through Google Health. Google Health is a platform through which the patient clinical information added by way of history and can share their profile Read more [...]
If you are looking for ways to measure social media, David Berkowitz has compiled 100 metric in this post: "100 Ways To Measure Social Media". The point is that would have to write another post entitled "100 ways to take advantage of the 100 metric" :-) ... I mean this is usually done a lot of 'trash measurement', not because it is poorly constructed metric, but because many are irrelevant metrics. The key is to have clear objectives that we set in the social media strategy and then clearly define how we will measure. Usually 3 or 4 KPI's should be sufficient to manage Read more [...]
For those who are more advanced with Google Analytics - Google today released new version of Google Analytics, with significant improvements -: - increasing to 20 the number of targets to be monitored - segmentation of unique visitors - ability to link with custom reports custom segments - intelligence & analytics analytics alerts - the pivot analysis allows secondary dimensions include ... You can see the entire list of news (here) Read more [...]
Microsoft Project Gemini is a new engine / datastore "in memory", enabling the publication of data for all tools that access the Analysis Services engine. Looks like we will soon and aims quite well, allowing to handle, for example from Excel using a plugin, large volumes of data very quickly. The combination with tools like Tableau, which is preparing for the new engine can be spectacular (and see). We will have to compare the performance vs. QlikView type tools, but certainly I think it will provide an important step in the development Read more [...]
Google just announced a couple of weeks to open the Google Analytics API. It is actually the API to access Google Analytics data monitors. This will wean developers and analysts of the current interface and enterprises to integrate with other information. It is a very important step forward in the Analytics Web environment and without doubt the possibilities to understand and optimize online activities. The documentation is in: Google Analytics API Read more [...]
Through the prism of the analyst, the visual analysis of information is one of the best techniques to understand the context that is analyzed in a fast, intuitive and quite often inconclusive. It is a promising route to early 90's driven by IBM and others, before being abandoned. Undoubtedly one of the best tools for viewing information on the market ... http://www.tableausoftware.com/
Using advanced modeling techniques allow us to establish the impact of marketing on the value of companies. At present the enterprise value is increasingly subject to aspects called 'intangibles' (simply for being off-balance sheet assets or difficult to measure). However, the customer base and its value are perfectly quantifiable aspects, and to measure the impact of marketing management in a "very tangible." If we can measure the time value of a customer and estimate the growth in customer numbers, we can calculate the future value of any company. Aspects Algunaos Read more [...]
How can we measure the success of the contents or investments Ads in advance? The 'Social Computing Lab', of HP Labs in Palo Alto, has developed a technique to measure and anticipate the popularity of content on YouTube or Digg inserted with a high degree of accuracy. Gabor Szabo and Bernardo A. Huberman have developed a model that basically looks at what happens in the first moments from the inclusion of content and from the first results, we get what the content popularity after 30 days. Such tools (early metrics) are critical Read more [...]
Stanford University lectures are reporting data mining course for those who can not attend their classes and for the enjoyment of those who have patience to be with some of the best specialists in the field. Recommended to view with time ahead ... last 1 hour.
I've always liked Kernel techniques like the use of wavelets for stock prediction. Within this broad spectrum of techniques based on kernels, I'm pretty impressed with the outcome of the approach of Professor Vapnik, certainly popularized this technique in academic settings and then served to the owner of Netflix to literally 'filthy rich'. Contrary to what you will find published, (for which you use Clementine documentation is painful), one of the propiedadades the use of this technique is low-input space dimensionality measurable or Read more [...]
The difference between the technical knowledge of data mining tools and construction of appropriate models and tools is similar to that between the last buy DIY toolkit in Leroy Merlin and build a house. When I look at the models being used in many companies focused on customer management: up-selling, cross-selling, churn prediction, segmentation ...; usually find the same type of faults in its construction, and do not enter the conceptual shortcomings, something I discuss in another post, Read more [...]
