During my time at LCG, I have been working in multiple tasks related to BP Europe, as the management of client activity and identification of changes in the customers’ consumption behaviour. This includes the development and upgrade of clients’ life cycle activities lists, which describe the different stages of consumption of a client: activation – clients begin to consume; rampup – clients start consuming and achieve their potential; prevention – customers have changed their behaviour and are now consuming less volume; retention – customers are about to leave the company, how can we prevent this. These are programmed in functions using the R language and delivered by loading of objects to BP’s salesforce and email notifications, via python-based applications.Data Mining and Analysis: Collecting data to answer to different types of requests related to the costumers’ consumption behaviour, building informative reports asked by the client by gathering and processing data that is stored in PostgreSQL databases, e.g volumes and card usages in a specific date and site.Automatization of manual processes: Improving their efficiency using ETL mechanisms for the ingestion and processing of data, as applying the multiple steps of the generation of life cycle activities lists to distinct country databases in sequence/parallel or transferring files from one place to another, using the Pentaho Data Integration tool.Machine learning-based predictive models: For example, based on a time series with the number of transactions received throughout time and a defined threshold, built a forecasting model that identifies when a received transactional file has a much lower number of transactions than expected and decides whether the information is updated to produce the life cycle activities lists. Jupyter notebooks with Python were used for the analysis of the data and processing of the input and the algorithms applied were from the R’s package ‘forecastHybrid’.

Achieved a grade of 18 (on a scale of 0 - 20).Systemic inhibition of angiogenesis (formation of blood vessels) is a clinically validated anticancer treatment strategy. However, the use of angiogenesis inhibitors to treat breast cancer often accelerates malignant tumor progression and metastasis. Increased intra-tumor heterogeneity (ITH) has been shown to emerge in renal cancer samples upon anti-angiogenesis treatment but the question remains whether increased ITH sponsors anti-angiogenesis therapy failure.To confirm that anti-angiogenic therapies drive ITH, tumors grown under “normal” or impaired supply of blood vessels were compared, using mouse xenograft models of breast cancer and the anti-angiogenic drug Bevacizumab (anti-VEGF). Genomic profiles were obtained by whole-exome sequencing of cancer cells from four distinct regions of the primary tumor, untreated or treated with Sunitinib. In each sample, mutations may have arisen, giving rise to new clones.In the first part of the practical work, the data gathered in these experiments was used to estimate the intra-tumor heterogeneity of the different tumors. The second part involved selecting the appropriate features from the sequence data and dimensionality reduction of gene mutations to study how these could distinguish between the two experimental groups.The Researcher also contributed and supported other team's initiatives related to the research project entitled “Targeting Intra-Tumor heterogeneity as a promising therapeutic strategy for cancer”.

Achieved a grade of 18 (on a scale of 0 - 20).The internship involved the following tasks:▪ Data analysis and reports production for external clients and other departments of the bank.▪ Tuning of previous storage techniques and modeling of big databases.▪ Automatic extraction of documents from SIBS.