# Fileset [RDA-Followup-NIMS-DICE.pdf](https://mdr.nims.go.jp/filesets/c8789232-3ad4-4fe0-b675-380bd7b9cc19/download) ## Creator [MATSUDA, Asahiko](https://orcid.org/0000-0001-5989-027X), [KADOHIRA, Takuya](https://orcid.org/0000-0003-0569-1309), [DEMURA, Masahiko](https://orcid.org/0000-0002-7308-3041) ## Rights Creative Commons BY Attribution 4.0 International[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/) ## Other metadata [NIMSʼs Materials Data Platform DICE and its future direction](https://mdr.nims.go.jp/datasets/0416a2af-484e-4cc5-b128-7578286abd53) ## Fulltext NIMSʼs Materials Data Platform DICEand its future directionNational Institute for Materials Science (NIMS)Research & Services Div. of Materials Data & Integrated System (MaDIS)https://orcid.org/0000-0001-5989-027XAsahiko Matsuda, Takuya Kadohira, Masahiko DemuraCC BY 4.0, except logos, stock images, or where otherwise notedhttps://orcid.org/0000-0001-5989-027Xhttps://creativecommons.org/licenses/by/4.0/2Re-introducing ourselves:• Five data-driven research groups• SIP-MI (Materials Integration) Laboratory• Materials Data Platform CenterSeven other research divisionsAbout 1600 peopleTsukuba Science City (~1hr from Tokyo)Research and Services Division of Materials Data and Integrated System© NIMS© NIMSFour main functionalities of DICECreatethe dataUsethe dataStorethe dataPublishthe dataExperiments &CalculationsAnalysis &IntegrationRepositoryStore and manageNIMS NOW International 17(1) 2019https://www.nims.go.jp/eng/publicity/nimsnow/2019/201901.html© NIMShttps://www.nims.go.jp/eng/publicity/nimsnow/2019/201901.html5NIMS NOW International 19(5) 2021 https://www.nims.go.jp/eng/publicity/nimsnow/2021/202105.html © NIMShttps://www.nims.go.jp/eng/publicity/nimsnow/2021/202105.htmlDICE 1.0 birds-eye viewCreatethe dataUsethe dataStorethe dataPublishthe dataFurther development for DICE 2.0, to spreadthis effort nationwide© NIMSCreating/Converting data7NIMS NOW International 19(5) 2021https://www.nims.go.jp/eng/publicity/nimsnow/2021/202105.htmlBinaries or other equipment-specific formats Plain-text based data© NIMShttps://www.nims.go.jp/eng/publicity/nimsnow/2021/202105.htmlAutomatic data collection from labsInstruments Offline PCWi-FiWi-Fi capableprogrammableSD cardM.Matsunami, A.Matsuda, T.Chikyow, Y.Harada, H.Yoshikawa, IPSJ Digital Practice 2(2) 80-89 (2021)https://doi.org/10.48505/nims.3050 (full text Japanese)Instruments Wired PCOne-way transfer Security devicehttps://doi.org/10.48505/nims.3050System and tools for measurement data collection and automatic conversion9binary raw datatext fileText conversion program with cooperation of the instrument manufacturersConverted numeric data(easier for humans)Python program (parser) that interprets the converted numeric data and visualizes themgraphProgram developed in-house, for format conversion and controlled metadataXMLmetadata“readable” dataNi3p“Schema-on-Read” data registration with user-customizable formatsSpectral data automatically sparse-modeledM. Suzuki et al., J. Vac. Sci. Technol. A 38 023204 (2020) https://doi.org/10.1116/1.5128408© NIMShttps://doi.org/10.1116/1.5128408Materials Data Conversion tools• Converts instrument-specific data format into readable formats• Plain-text formats and graphs• Converters were developed through cooperation of instrument manufacturers• XPS (survey, narrow, depth) by ULVAC-PHI• XRD, 2D-XRD by Rigaku• AES by JEOL• AES (survey, depth) by ULVAC-PHI• github.com/nims-dpfc• Runs on Jupyter Notebook • Wouldnʼt it be nice if…• we start doing this more? Convert all of your data in a stable, readable format• we can get more techniques from more manufacturers on board?10Storing/Searching data: Metadata• Four different styles, even within the Center (not mutually exclusive)• Multi-layer nested grand metadata schema• Simplified metadata with a focus on data repository• Metadata as directory names• Metadata as plain-text files11Matsuda et al. “Materials metadata: as a custom schema, as directories, or in a data package”RDA Virtual Plenary 15 (2020) https://doi.org/10.48505/nims.3031https://doi.org/10.48505/nims.303112DICE Common Message Format 1.0Characterization metadataMethod,Environment…Specimen metadataMaterial type,Structure…PropertymetadataPhysical properties,Units…Synthesis/ProcessmetadataProcessed date,Temperature…CalculationmetadataComputer software,Version…Characterizationprimary paramsSpecimen primary paramsProperty primary paramsSynthesis/Processprimary paramsCalculationprimary paramsData DataData Data DataMandatory metadataDomain-specific metadataPrimary parametersImplementedas data modelSave as filesMETADATADATABibliographic metadata Administrative metadata Subject material++S. Kikuchi, T. Kadohira, M. Suzuki, H. Naito, IEICE Tech. Rep. 119 SC2019-2 Full text (Japanese): https://mdr.nims.go.jp/concern/publications/m039k565zSchema: https://doi.org/10.48505/nims.3240• JSON Schema• System-to-system communication✔ Highly descriptive, high reusability. ✘ Quite complex: Not human-readable.Implementation difficulties.https://mdr.nims.go.jp/concern/publications/m039k565zhttps://doi.org/10.48505/nims.3240Vocabulary-based metadata transfer between systemsData Collection Message PublishRDE local dictionarysyncedAPI queryreferMatVoc ID: Q386M. Ishii, H. Nagao, A. Matsuda, K. Tanabe, H. Yoshikawa, 23rd XAFS Forum (2020) (Japanese)MaterialsVocabularyWikiMDR Schema 2.0• Single-layer• no multi-level nesting• YAML ✔ no brackets!✔ in-file comments!• Deliberately simple (relatively)• focus on repository use• Support for common IDs• Can be validated with “Yamale”14(snippet)github.com/nims-dpfc/mdr-schemadoi: 10.48505/nims.3239https://github.com/nims-dpfc/mdr-schemahttps://doi.org/10.48505/nims.3239Metadata as directories15instruments & lab PCs📂 OSC-100 (Instrument ID)└📂 u01234 (User Name/ID)└📂 ProjABC (Project Name)└📄 data.csvMapped to appropriatemetadata fieldsInstrument, User, Project...platform systemsImplemented as part of our IoT-assisted data collection system.Pre-defined structure:✔ Easy to use in labs.✔ Network not required.✘ Prone to human errors.✘ Only simple common metadata.✘ Different mapping for every research project.RDA Virtual Plenary 15 https://doi.org/10.48505/nims.3031https://doi.org/10.48505/nims.3031Metadata in a file / data package16root directory: /ro-crate-metadata.jsonJSON-LD mandatory metadata{}ro-crate-preview.htmlHTML preview for humans (optional)<>ro-crate-preview_files/Files for HTML Payload datahttps://www.researchobject.org/ro-crate/K.Tanabe & A.Matsuda, Joho Chishiki Gakkaishi 31(4) 497-502 (2021)https://doi.org/10.2964/jsik_2021_067 (full text Japanese)https://www.researchobject.org/ro-crate/https://doi.org/10.2964/jsik_2021_067JSON-LD metadata for RO-Crate package17Package infoMaterial/SubstanceDataset descriptionDataAdditional metadata file referenceScientific metadata(test environment)K.Tanabe & A.Matsuda, Joho Chishiki Gakkaishi 31(4) 497-502 (2021)https://doi.org/10.2964/jsik_2021_067 (full text Japanese)https://doi.org/10.2964/jsik_2021_067Metadata in a file: Some notes• RO-Crateʼs main metadata file is JSON-LD• … but the package can include any other file as metadata• e.g., aforementioned MDR Schema YAML• A package doesnʼt have to be RO-Crate• … but at the least, itʼs nice to base it on plain text✔ Files are relatively stable, separated from a system.✘ Will not be easily searchable on a system. What are the essential metadata we want across various data?(Iʼm far from the first person mentioning this question, but this is still a key issue from my view!)18Using the data19• In-silico link among process, structure, property, performance• Integrate experiments and computations by data driven approachesHeat-treatment Image analysis Mechanical testsMaterials Integration by Network TechnologyC Si Mn Mo …… Al ……mass % mass % mass % mass % …… mass % ……炭素 珪素 Mo Mn …… Al-sol. Al-insol. Al-total ……at % at % at % at % …… at % at % at % ……Absorb differences in wording, order, units, collection items, etc.Unit Wording Order FinenessData schema based on “dimension of data structure”Heat C Si MnMBB 0.12 0.36 0.52MBC 0.09 0.37 0.49MBD 0.1 0.28 0.49MBE 0.12 0.43 0.48Heat Temp. 0.2％PS TSMBB 23 342 490MBB 100 338 454MBB 200 337 465MBB 300 346 495Heat Temp. Stress Rupture timeMBB 500 373 712.7MBB 500 333 2221.9MBB 500 294 5941.6MBB 500 265 11821.6Materials info.: One dim. Tensile property: Two dim. Creep property: Three dim.Reference datasheet• NIMS creep/fatigue datasheet• ISIJ datasheet• JAEA datasheetXSDWeʼve designed data structure for our materials problem, which will be published.Demura, RDA P18Wouldn't it be in the public interest to increase the number of data schema?Share well-defined data schemaExample of designed data structure© NIMSSharing data structures• Data can be made available and readable• But thereʼs a barrier in actually reading (understanding what they represent) and utilizing data20Two data about the same topic (supposedly)• How can we be sure that we can mash up these two data?• What would tell me that these two columns are directly comparable?Attach data structure descriptions to all data and share them somehow? Ideas?Conclusion• We built “DICE” as a platform where our researchers can create, store, use, and publish their data (the four actions of the platform), supported through various subsystems• There are many areas where we all have benefited from and will benefit from continued discussions and cooperation:• Making it easier to store data in a stable manner• Metadata, how and what• Describing and sharing data structure• and probably more…21Further reading• NIMS NOW International• MaDIS special issue (2019) https://www.nims.go.jp/eng/publicity/nimsnow/2019/201901.html• Materials DX special issue (2021) https://www.nims.go.jp/eng/publicity/nimsnow/2021/202105.html• M. Tanifuji, A. Matsuda, H. Yoshikawa: Materials Data Platform - a FAIR System for Data-Driven Materials Science. Proc. 2019 8th Int. Congress on Advanced Applied Informatics (IIAI-AAI). 2019• T. Takemura, M. Ishii, M. Tanifuji: PoLyInfo RDF: A Semantically Reinforced Polymer Database for Materials Informatics. CEUR Workshop Proceedings. 2019, 69-72.• L. Foppiano, L. Romary, M. Ishii, M. Tanifuji: Automatic Identification and Normalisation of Physical Measurements in Scientific Literature. DocEng '19 Proc. ACM Symp. on Document Eng 2019.• M. Suzuki, H. Nagao, Y. Harada, H. Shinotsuka, K. Watanabe, A. Sasaki, A. Matsuda, K. Kimoto, H. Yoshikawa: Raw-to-Repository Characterization Data Conversion for Repeatable, Replicable, and Reproducible Measurements. J. Vac. Sci. Technol. A 38 [2] 023204 (2020).• X. Liu, L. Yang, Z. Hou, Bo Da, K. Nagata, H. Yoshikawa, S. Tanuma, Y. Sun, Z. Ding: Machine learning approach for the prediction of electron inelastic mean free paths. Phys. Rev. Mater. 5 [3] (2021).• H. Oka, A. Yoshizawa, H. Shindo, Y. Matsumoto, M. Ishii: Machine extraction of polymer data from tables using XML versions of scientific articles. Sci. Technol. Adv. Mater. Methods 1, 12 (2021).• S. Dieb, K. Amano, K. Tanabe, D. Sato, M. Ishii, M. Tanifuji: Creating Research Topic Map for NIMS SAMURAI Database Using Natural Language Processing Approach. Sci. Technol. Adv. Mater. Methods 1, 2 (2021).22https://www.nims.go.jp/eng/publicity/nimsnow/2019/201901.htmlhttps://www.nims.go.jp/eng/publicity/nimsnow/2021/202105.htmlhttps://doi.org/10.1109/IIAI-AAI.2019.00206http://ceur-ws.org/Vol-2456/paper18.pdfhttps://doi.org/10.1145/3342558.3345411https://doi.org/10.1116/1.5128408https://doi.org/10.1103/PhysRevMaterials.5.033802https://doi.org/10.1080/27660400.2021.1899456https://doi.org/10.1080/27660400.2021.1899426