70 likes | 85 Views
Learn about the innovative data analysis tools developed by Dr. Ramona Lichtenthäler for identifying early risks in food controls. This includes a nationwide database for optimizing information exchange and a system for automating risk identification. Explore the implementation process and the main project goals at the 32nd Meeting of the Forum of Food Law Enforcement Practitioners in Berlin.
E N D
Data analysis tools for early risk identification Dr. Ramona Lichtenthäler
History and Implementation Data analysis within the pilot project eFI 32nd Meeting of the Forum of Food Law Enforcement Practitioners Berlin, Dr. Ramona Lichtenthäler • The pilot project “electronic early identification and information system”, in the scope of Reg. 882/2004/EC is a result of considerations following the spoiled meat scandals in Germany (2005 ff): • An early identification system to scan risks and trends in controls data as well as a geo-referenced visualisation of data have been identified to be adequate tools to optimise food control. • A nationwide database would improve information exchange between the respective authorities. • For implementation of both, a working group of the Federal states and the Federal government was built by order of the Consumer Protection Ministerial Conference (VSMK).
Main project goal = risk identification in controls data 32nd Meeting of the Forum of Food Law Enforcement Practitioners Berlin, Dr. Ramona Lichtenthäler • Goal: • To early identify risks and trends of products and establishments • To identify hazards, before they become a crisis • To show general developments and tendencies („trends“) • Approach: • Nationwide consolidation and analysis of data indicate relationships, which wouldn‘t be apparent otherwise • Consolidation of sample analysis data and establishment inspection data leads to new conclusions • Challenge: identification of relevant data
= risk identification in controls data Main project goal 32nd Meeting of the Forum of Food Law Enforcement Practitioners Berlin, Dr. Ramona Lichtenthäler • Technical implementation • Data is analysed regularly and automatically, e. g.: • Comparison with reference distribution (mean value, shape and width of the distribution, …) • Comparison with maximum values and other reference values (limit values, signal values…) • Establishment of time series (new risks, „retro risks“, percentage change, data gaps, …) • Identification of conformities (correlations, pattern recognition,…) • Potentially interesting results are filtered out and classified into risk levels (information, warning, alert) • System informs responsible operator by e-mail if risk level is reached
Infrastructure and output For data analysis an updated data basis is needed Data transmission Data analysis Data collection eFI Data delivery web portal Nation data-base BI-Software Analysis-Software Analysis resultsandreports Länder-specific applications (z.B. LIMS or BALVI-server) ? interfaces eFI- analysis/reporting web portal Data of sample analysis Controls data of inspected establishments Data selection, Anonymisation, Transformation, Forwarding Concept of user rights and rolls, Anonymisation, Data selection, Data analysis, Data reporting Plausibility check 32nd Meeting of the Forum of Food Law Enforcement Practitioners Berlin, Dr. Ramona Lichtenthäler
Analysis results and reports Data plottet on maps Chart report Are topic-associated .. Tile charts of all topic associated data Evaluation of one sample Trend visualisation Exhaustion of max. value Reg. max. value max. value Exhaustion mean of mean (5 years) Exhaustion of max. value Higher exhaustion of maximum values …and will be discussed with the Laender experts to evaluate the relevance and to take action if applicable. 32nd Meeting of the Forum of Food Law Enforcement Practitioners Berlin, Dr. Ramona Lichtenthäler