Just focus on decisions !
DEXTER™ is simply an agile & pragmatic data wrangling tool. Our goal is to offer advanced features just by clicking in a menu, no Data Sciences degree required! The very first feature in DEXTER™ is to streamline and clean the different formats encountered in data bases (e.g. French & english date formats, incorrect manual encoding, missing values, “outliers” to define and then transform). DEXTER™ will allow you to enter in the ECO ANALYTICS epoch.
An easy software, practical and stand alone to wrangle your data with high “dexterity”: this made us to christen it DEXTER!
DEXTER permet d’effectuer des créations de champs (ex: troncature, raccourcissement, extensions markoviennes) et offre des techniques de représentations des données élémentaires (nuages, corrélations) ou plus évoluées (projection par t-SNE, binnage topologique auto-organisé). Des rapports analytiques sont également disponibles : statisques élémentaires ou par “regroupements” amont.
DEXTER allows you to create new fields (e.g. truncature or shortening, Markovian extension). It offers data visualization techniques, rudimentary (XY-chart) or more advanced (t-Stochastic Neighbourhood Embedding or topological binning). Direct analytic reports are also available, such as elementary statistics. Five sampling methods can be used (such as R or R-fast reservoir) in order to extract relevant sub-data bases. All the data sets (imported, or created) are viewed as nodes in a graph, as for reports and any visual representations. For users interested in stochastic approaches, it will be possible by end of 2016 to ‘micmic’ data set by using probabilistic Haar wavelet decompositions, and to generate stochastic statistics report.
We made the choice in DEXTER™ not to allow manual modification: any action on data require to trigger an algorithm (*) !
Our software is a tiny jewel of simplicity that allows to work on serious-sized set of data. For illustration, few hours are needed to convert in DEXTER format a data base of hundreds of millions of records, each records containing hundreds of fields! But since we follow theECO-ANALYTICS, principle, we have no need to hurry, no?
DEXTER™ is an autonomous software that works without data base system; furthermore, there is no saving manual process, everything is steered by DEXTER™! Our software adapts also to the power and the memory of your machine.
DEXTER™ is distributed in a fermium mode.
Nota Bene: for companies that need a more industrial approach, in 2017 we will propose DEXTER+™, that works on grapes of lap-top with distributed tasks, and that propose also a NO SQL data base component (from our partner QUASAR DB).
(*): come on, let’s tell you the true: you can manually one data at once by using a GUI voluntary… crabbed 😉