Real-time Products
Impact Analysis
To determine whether a user, brand or a topic has some impact is extremly important. A good measure allows you to identify important parts in social communication. That is why we have started TWIMPACT. We have created algorithms that calculate sustainable impact scores. A number that tells you whether the subject has sustainable influence on its environment over time. The "twimpact" is a number that can easily go up but over time it requires the subject to continually influence to keep it up or get higher.
To ensure that we are not just doing simple statistics on follower counts or retweets, we actually look into the structure of the communication that is often not represented in simple machine readable ways. This includes mentions mentions classic retweets.
Watch the global impact analysis results on our demo site.
Stream Trending
Social networks emit millions of parallel conversations. It is an ocean with big streams that we follow in real-time to extract trends for selected topics in real-time. You select what you want to look out for and we will present the trends as they evolve without delay including, but not limited to, influential users, tag, link or media trends.
In order to find relevant information on social network platforms, you have to be able to discover topics and users relevant for you in real-time. Just searching for matching updates in real-time doesn't solve the problem, because recency is only a weak indicator for relevance. Trending solves this problem by analyzing in real-time which updates, links, keywords or hashtags are mentioned significantly more often than the rest to identify interesting pieces of information. Twimpact analyzes information streams in real-time and lets you zoom in on current events whenever you want to slow down and inspect what is happening.
Custom Data Correlation
While extracting the data from social network streams is an interesting topic itself, we understand that you may want to associate and relate the results to your own data feed to detect correlated trends and effects. We provide the infrastructure for customized real-time trending to achieve just that: discovering interesting pieces of information in high-volume real-time streams.
Why customized?
Given the sheer amount of data on social networks like Twitter, just looking at global trends will only return what the largest group of users is interested in. Looking at the trending topics of Twitter is a good example. It is filled with generic hashtags like #damnitstrue or big events. Actually, many different kinds of topics are still in the stream, only on a smaller scale. During the revolution in Egypt, topics related to Egypt seldom trended, although explicitly filtering for #egypt and related keywords revealed a whole stream of interesting and relevant updates.
Why real-time?
Given the rate at which topics unfold, there you cannot rely on static analyses of the past. Even if you do those in regular intervals, you are always struggling to catch up with current trends. Therefore, the analysis needs to be in real-time.
Research & Development
We are constantly working on new algorithms and methods to extract relevant information from data streams. Twimpact models modern machine learning-based ranking that do not rely on simple reader/follower counts, but rather includes actual communication patterns, such as citations in publications or retweets from Twitter. This allows us to provide you with factual information on reach and impact for each group and individual taking part in online conversations about your market and products.
With our close connection to one of the leading research labs in data analysis at the Berlin Institute of Technology we are able to apply the latest mathematical developments to our data analysis efforts.