In our continuous development of MESLIS WebApp during the last couple of weeks we set a strong focus on artist detection, event tags, data matching and key performance indicators.
MESLIS clients, collecting societies worldwide often need to know what artists performed on particular events to define the right license and correct amount of invoice for licensing those events. Some collecting societies even provide scores to individual artists to rank them. The web content extraction feature of MESLIS from now on automatically identifies and extracts all artist names from any digital, text based event description. These artist names are listed for sorting, searching, filtering and processing within a certain column and field in MESLIS WebApp. All users can search and filter for those artists and even select from an individually defined range of artist scores.
Picture: Search for artist within a certain range of score
Event tags have a similar importance to collecting societies. MESLIS classification models are now able to automatically identify what particular type of event and music usage is described in individual event advertisements. A necessary prerequisite for this is a common definition of required event tags individually with every single interested collecting society and providing good training data to learn MESLIS classification model. If this is fulfilled, the brand new event tag feature gives the customer a very good opportunity to search and filter specifically for live music events, DJ music events or others.
Picture: Detail view with matched event IDs and event tags
Matching MESLIS data with customer data is one of the most important features of our Music-Event-System-Location-Ident-System. Only after matching the data the MESLIS user is able to sort out those locations and events which his / her collecting society does not yet know and has not yet sold a license for. This unmatched (unknown) data provides the opportunity for additional licensing, additional revenue, an increase of royalties for rights holders and an increase of market coverage. MESLIS WebApp provides customers IDs on events and locations (after matching) in specific columns and fields. Using the filter options, the MESLIS user can easily search for matched and unmatched (in fact unknown / new to customer) data.
Picture: Matched event IDs in table view
MESLIS is a self learning software (machine learning). It automatically determines the ranking of locations and the relevance of events to the customer according to many different criteria. The quality and accuracy of scoring, ranking and relevance assessment improves automatically over time (self learning). But the self learning features only work good as long as they have good training data to learn from. The one who still knows best which locations are most and which are less valuable or what events are most, less or even not at all relevant is the customer himself. Therefore we have implemented specific features in MESLIS on the basis of which the customer produces training data automatically once he / she starts working with MESLIS. All processed or exported data is considered as positive training data and all scrapt data will be taken as negative training data.
Defining and measuring several key performance indicators (KPIs) is crucial for continuously managing performance, efficiency and success of using MESLIS. deecoob provides some basic dashboards to every single customer. Those dashboards for instance show quantity of raw data and matched data, quality of classification model, the course of data crawling, allocation of events on a timescale, efficiency of MESLIS workflow and many more KPIs. All users have now access to these dashboards directly from MESLIS WebApp.
