For a long time there has been issues about time estimation in IT-projects. Despite the fact that we went from the traditional waterfall model to the agile methods, the challenges of time estimation has continued. The agile methods do not solve the problems. A technique that agile teams have chosen to apply in order to solve this problem is planning poker. The technique is based on a method that includes a number of steps for implementation. This study is conducted to investigate whether methods such as planning poker for time estimation provide agile team support to make safer time estimates in agile IT-projects. The study also investigates how the process of planning poker is applied in practice. The study is based on interviews and a literature study that has compiled in a result that has been analysed. The result shows that agile teams choose to modify the prescribed method of planning poker to streamline the process as they find that some steps in the prescribed method are unnecessary. This result differences from earlier research where praxis is that you follow all steps of the prescribed method. Further a number of challenges that comes at an adaptation of the technique are identified. The result also shows that planning poker creates better discussion and understanding of matters for the agile team.

Forms are used in many situations. For example, they tend to be ubiquitous in communications between individuals and government agencies. Something which could potentially boost transparency and efficiency is accurate estimates of how long it will take for the receiver to process a given completed form. Unfortunately, such estimates are often not available. This thesis examines the problem of using machine learning to predict form processing times, applied to the context of Swedish pay-as-you-earn tax returns. More specifically, it compares a naive baseline model to several random forest models, some based on the more common batch learning principle, and others on online learning which is typically seen as more suitable for working with data streams and changing conditions. Despite the theoretical advantages of online learning, none of the models using that approach were able to consistently outperform the naive baseline model. Conversely, the two primarily evaluated batch learning models were successful in doing so, although the improvement over the baseline was small. / Formulär används i många sammanhang. De är exempelvis mycket vanligt förekommande i kommunikation mellan privatpersoner och myndigheter. Något som potentiellt skulle kunna innebära ökad transparens och effektivitet är träffsäkra uppskattningar av hur lång tid det tar för mottagaren att handlägga ett givet formulär. Dessvärre är sådana uppskattningar ofta inte tillgängliga. Detta examensarbete undersöker hur maskininlärning kan användas för att förutspå formulärhandläggningstider, tillämpat i kontexten svenska arbetsgivardeklarationer. Mer specifikt jämförs en enkel naiv modell mot flera random forest-modeller, vissa baserade på den vanligare batchinlärningsprincipen, och andra på onlineinlärning som brukar ses som mer passande för dataströmmar och föränderliga förhållanden. Trots de teoretiska fördelarna med onlineinlärning lyckades inte någon av modellerna som använde sig av den tekniken konsekvent ge bättre resultat än den naiva grundmodellen. Däremot visade sig de två primärt utvärderade batchinlärningsmodellerna vara framgångsrika i det avseendet, även om skillnaden mot den naiva modellen var liten.