StatAILab/DataSciEval
2
1<!DOCTYPE html>2<html lang="en">3<head>4 <meta charset="UTF-8">5 <meta name="viewport" content="width=device-width, initial-scale=1.0">6 <meta name="description" content="DataSciEval evaluates LLMs and agents across data science theory and methods and end-to-end real-world data analysis.">7 <meta name="keywords" content="DataSciEval, data science, benchmark, LLM, agent, data science knowledge, machine learning, artificial intelligence">8 <title>DataSciEval | Data Science Benchmark for LLMs and Agents</title>9 <style>10 :root {11 --ink: #14213d;12 --muted: #5e6b82;13 --line: #dfe5ed;14 --paper: #ffffff;15 --soft: #f5f6f4;16 --navy: #17324d;17 --navy-2: #234f68;18 --yellow: #f7c948;19 --yellow-soft: #fff5cf;20 --purple: #7057d9;21 --purple-soft: #f0edff;22 --teal: #087f7a;23 --teal-soft: #e6f7f5;24 --blue: #2869d8;25 --red: #c84858;26 --spectrum-1: #7c5cff;27 --spectrum-2: #367bf5;28 --spectrum-3: #00a6a6;29 --spectrum-4: #2fbf71;30 --spectrum-5: #f3c64e;31 --spectrum-6: #ef7b45;32 --shadow: 0 2px 8px rgba(20, 33, 61, 0.05);33 --shadow-sm: 0 1px 4px rgba(20, 33, 61, 0.05);34 }35 36 * {37 box-sizing: border-box;38 }39 40 html {41 scroll-behavior: smooth;42 -webkit-text-size-adjust: 100%;43 text-size-adjust: 100%;44 }45 46 body {47 margin: 0;48 max-width: 100%;49 overflow-x: hidden;50 color: var(--ink);51 background: #fff;52 font-family: Inter, ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif;53 line-height: 1.6;54 }55 56 a {57 color: inherit;58 text-decoration: none;59 }60 61 button {62 font: inherit;63 }64 65 img {66 display: block;67 max-width: 100%;68 height: auto;69 }70 71 section[id],72 #datasets {73 scroll-margin-top: 82px;74 }75 76 :focus-visible {77 outline: 3px solid rgba(40, 105, 216, 0.45);78 outline-offset: 3px;79 }80 81 .shell {82 width: min(1180px, calc(100% - 40px));83 margin: 0 auto;84 }85 86 .hero {87 position: relative;88 overflow: hidden;89 color: var(--ink);90 border-bottom: 1px solid var(--line);91 background: #fff;92 }93 94 .hero::before {95 position: absolute;96 inset: 0;97 display: none;98 }99 100 .hero::after {101 position: absolute;102 top: 0;103 right: 0;104 bottom: auto;105 left: 0;106 width: 100%;107 height: 5px;108 border-radius: 0;109 display: none;110 background: var(--navy);111 content: "";112 filter: none;113 opacity: 0.9;114 transform: none;115 }116 117 .hero-inner {118 position: relative;119 z-index: 1;120 display: block;121 min-height: auto;122 padding: 72px 0 48px;123 text-align: center;124 }125 126 .hero-inner > *,127 .shell > * {128 min-width: 0;129 }130 131 .hero h1,132 .hero-subtitle,133 .hero-copy {134 width: 100%;135 max-width: 100%;136 overflow-wrap: anywhere;137 }138 139 .eyebrow {140 display: inline-flex;141 align-items: center;142 gap: 9px;143 margin: 0 0 20px;144 color: var(--blue);145 font-size: 13px;146 font-weight: 800;147 letter-spacing: 0;148 text-transform: uppercase;149 white-space: normal;150 }151 152 .eyebrow::before {153 width: 28px;154 height: 2px;155 background: var(--blue);156 content: "";157 }158 159 h1,160 h2,161 p {162 margin-top: 0;163 }164 165 h1,166 h2,167 h3 {168 font-family: Georgia, "Times New Roman", serif;169 }170 171 h1 {172 max-width: 1040px;173 margin-right: auto;174 margin-left: auto;175 margin-bottom: 20px;176 font-size: 82px;177 line-height: 0.98;178 letter-spacing: 0;179 }180 181 h1 span {182 color: inherit;183 background: none;184 }185 186 .hero-subtitle {187 max-width: 1080px;188 margin-right: auto;189 margin-left: auto;190 margin-bottom: 15px;191 color: #33445a;192 font-size: 24px;193 font-weight: 650;194 line-height: 1.3;195 }196 197 .hero-copy {198 max-width: 760px;199 margin-right: auto;200 margin-bottom: 28px;201 margin-left: auto;202 color: var(--muted);203 font-size: 17px;204 }205 206 .hero-summary {207 max-width: 980px;208 margin: 30px auto 0;209 padding-top: 26px;210 border-top: 1px solid var(--line);211 color: #4f5f75;212 font-size: 15px;213 line-height: 1.75;214 }215 216 .hero-summary strong {217 color: var(--navy);218 font-weight: 850;219 }220 221 .hero-summary a {222 color: var(--blue);223 font-weight: 800;224 }225 226 .hero-summary a:hover {227 text-decoration: underline;228 }229 230 .hero-datasets {231 display: grid;232 grid-template-columns: 170px repeat(2, minmax(0, 1fr));233 gap: 10px;234 max-width: 980px;235 margin: 22px auto 0;236 padding: 12px;237 border: 1px solid var(--line);238 border-radius: 4px;239 background: #f8fafc;240 box-shadow: var(--shadow-sm);241 text-align: left;242 }243 244 .hero-datasets-intro {245 display: flex;246 flex-direction: column;247 justify-content: center;248 padding: 4px 10px;249 }250 251 .hero-datasets-intro span {252 color: var(--blue);253 font-size: 10px;254 font-weight: 900;255 letter-spacing: 0.1em;256 text-transform: uppercase;257 }258 259 .hero-datasets-intro strong {260 color: var(--navy);261 font-size: 14px;262 }263 264 .hero-data-link {265 position: relative;266 display: grid;267 grid-template-columns: 1fr auto;268 gap: 2px 14px;269 align-items: center;270 min-height: 78px;271 padding: 12px 14px;272 border: 1px solid var(--line);273 border-radius: 4px;274 background: #fff;275 transition: transform 150ms ease, border-color 150ms ease, box-shadow 150ms ease;276 }277 278 .hero-data-link:hover {279 transform: translateY(-2px);280 box-shadow: 0 7px 18px rgba(20, 33, 61, 0.1);281 }282 283 .hero-data-link:focus-visible {284 outline: 3px solid rgba(40, 105, 216, 0.28);285 outline-offset: 3px;286 }287 288 .hero-data-link .data-track,289 .hero-data-link strong,290 .hero-data-link small {291 display: block;292 grid-column: 1;293 }294 295 .hero-data-link .data-track {296 font-size: 10px;297 font-weight: 900;298 letter-spacing: 0.08em;299 text-transform: uppercase;300 }301 302 .hero-data-link strong {303 color: var(--ink);304 font-size: 15px;305 }306 307 .hero-data-link small {308 color: var(--muted);309 font-size: 11px;310 font-weight: 650;311 }312 313 .hero-data-link .download-arrow {314 grid-column: 2;315 grid-row: 1 / 4;316 display: inline-flex;317 width: 34px;318 height: 34px;319 align-items: center;320 justify-content: center;321 border-radius: 50%;322 color: #fff;323 font-size: 19px;324 }325 326 .hero-data-link.theory-data {327 border-left: 4px solid var(--purple);328 }329 330 .hero-data-link.theory-data .data-track {331 color: var(--purple);332 }333 334 .hero-data-link.theory-data .download-arrow {335 background: var(--purple);336 }337 338 .hero-data-link.application-data {339 border-left: 4px solid var(--teal);340 }341 342 .hero-data-link.application-data .data-track {343 color: var(--teal);344 }345 346 .hero-data-link.application-data .download-arrow {347 background: var(--teal);348 }349 350 .actions,351 .nav-links,352 .tag-row,353 .tabs {354 display: flex;355 flex-wrap: wrap;356 gap: 10px;357 justify-content: center;358 }359 360 .button {361 display: inline-flex;362 align-items: center;363 justify-content: center;364 min-height: 46px;365 padding: 10px 17px;366 border: 1px solid #b8c3cf;367 border-radius: 4px;368 color: var(--ink);369 background: #fff;370 font-size: 14px;371 font-weight: 780;372 transition: transform 150ms ease, background 150ms ease, border-color 150ms ease;373 }374 375 .button:hover {376 transform: translateY(-2px);377 border-color: var(--blue);378 background: #f3f7fb;379 }380 381 .button.primary {382 border-color: var(--navy);383 background: var(--navy);384 color: #fff;385 }386 387 .button.primary:hover {388 background: #244866;389 }390 391 .button.dark {392 border-color: var(--navy);393 background: var(--navy);394 color: #fff;395 }396 397 .button.outline {398 border-color: var(--line);399 color: var(--ink);400 background: #fff;401 }402 403 .hero-map {404 display: none;405 }406 407 .hero-scope {408 display: grid;409 grid-template-columns: repeat(2, minmax(0, 1fr));410 max-width: 820px;411 margin-top: 34px;412 margin-right: auto;413 margin-left: auto;414 border-top: 1px solid #cfd7df;415 border-bottom: 1px solid #cfd7df;416 }417 418 .hero-scope {419 display: none;420 }421 422 .hero-scope div {423 padding: 16px 18px;424 border-right: 1px solid #cfd7df;425 }426 427 .hero-scope div:last-child {428 border-right: 0;429 }430 431 .hero-scope strong {432 display: block;433 margin-bottom: 3px;434 color: var(--ink);435 font-family: Georgia, "Times New Roman", serif;436 font-size: 17px;437 }438 439 .hero-scope span {440 color: var(--muted);441 font-size: 12px;442 }443 444 .orbit {445 display: none;446 }447 448 .orbit.two {449 display: none;450 }451 452 .core {453 position: static;454 display: block;455 width: auto;456 height: auto;457 padding-bottom: 15px;458 border: 0;459 border-bottom: 1px solid var(--line);460 border-radius: 0;461 background: transparent;462 box-shadow: none;463 text-align: left;464 transform: none;465 }466 467 .core strong {468 display: block;469 color: transparent;470 background: none;471 background-clip: text;472 -webkit-background-clip: text;473 font-size: 30px;474 line-height: 1;475 letter-spacing: 0;476 transform: none;477 }478 479 .core span {480 display: block;481 margin-top: 6px;482 color: var(--muted);483 font-size: 12px;484 font-weight: 800;485 letter-spacing: 0;486 text-transform: uppercase;487 }488 489 .orbit-card {490 position: static;491 width: auto;492 padding: 16px;493 border: 1px solid var(--line);494 border-radius: 3px;495 background: #fafbfc;496 box-shadow: none;497 }498 499 .orbit-card.theory {500 border-top: 3px solid var(--spectrum-1);501 }502 503 .orbit-card.practice {504 border-top: 3px solid var(--spectrum-4);505 }506 507 .orbit-card .mini-label {508 margin-bottom: 5px;509 color: var(--muted);510 font-size: 11px;511 font-weight: 800;512 letter-spacing: 0;513 text-transform: uppercase;514 }515 516 .orbit-card strong {517 display: block;518 margin-bottom: 5px;519 font-size: 18px;520 }521 522 .orbit-card p {523 margin: 0;524 color: var(--muted);525 font-size: 13px;526 line-height: 1.45;527 }528 529 .axis-chip {530 display: none;531 }532 533 .axis-chip.one {534 top: 86px;535 right: 5px;536 }537 538 .axis-chip.two {539 bottom: 80px;540 left: 20px;541 }542 543 .nav {544 position: sticky;545 top: 0;546 z-index: 20;547 border-bottom: 1px solid rgba(20, 33, 61, 0.10);548 background: rgba(251, 251, 250, 0.97);549 -webkit-backdrop-filter: blur(14px);550 backdrop-filter: blur(14px);551 }552 553 .nav .shell {554 display: flex;555 min-height: 62px;556 align-items: center;557 justify-content: space-between;558 }559 560 .brand {561 display: inline-flex;562 gap: 10px;563 align-items: center;564 color: var(--navy);565 font-weight: 900;566 letter-spacing: 0;567 }568 569 .brand-mark {570 display: grid;571 width: 31px;572 height: 31px;573 border-radius: 3px;574 color: #fff;575 background: var(--navy);576 font-size: 11px;577 place-items: center;578 }579 580 .nav-links {581 gap: 4px;582 }583 584 .nav-links a {585 padding: 8px 10px;586 border-radius: 7px;587 color: var(--muted);588 font-size: 13px;589 font-weight: 750;590 }591 592 .nav-links a:hover {593 background: var(--soft);594 color: var(--ink);595 }596 597 section {598 padding: 70px 0;599 }600 601 section.soft {602 background: var(--soft);603 }604 605 main {606 display: flex;607 flex-direction: column;608 }609 610 main > .stats {611 order: 0;612 }613 614 #overview {615 order: 1;616 }617 618 #leaderboard {619 order: 2;620 }621 622 #tracks {623 order: 3;624 }625 626 #examples {627 order: 4;628 }629 630 #evaluation {631 order: 5;632 }633 634 #reference {635 order: 6;636 }637 638 .section-kicker {639 margin-bottom: 9px;640 color: var(--blue);641 font-size: 12px;642 font-weight: 850;643 letter-spacing: 0;644 text-transform: uppercase;645 }646 647 .section-kicker,648 .track-id,649 .track-name,650 .tag-row {651 display: none;652 }653 654 .section-head {655 display: flex;656 gap: 24px;657 align-items: end;658 justify-content: space-between;659 margin-bottom: 30px;660 }661 662 h2 {663 max-width: 820px;664 margin-bottom: 10px;665 font-size: 40px;666 line-height: 1.1;667 letter-spacing: 0;668 }669 670 h3 {671 line-height: 1.25;672 }673 674 .section-head p,675 .section-copy {676 max-width: 790px;677 margin-bottom: 0;678 color: var(--muted);679 font-size: 16px;680 }681 682 .stats {683 display: grid;684 grid-template-columns: repeat(4, minmax(0, 1fr));685 gap: 14px;686 margin-top: 28px;687 margin-bottom: 0;688 position: relative;689 z-index: 3;690 }691 692 .stats {693 display: none;694 }695 696 .stat {697 padding: 23px;698 border: 1px solid var(--line);699 border-radius: 3px;700 background: #fff;701 box-shadow: none;702 }703 704 .stat strong {705 display: block;706 margin-bottom: 4px;707 color: var(--navy);708 font-size: 29px;709 line-height: 1;710 letter-spacing: 0;711 }712 713 .stat span {714 color: var(--muted);715 font-size: 13px;716 font-weight: 650;717 }718 719 .thesis {720 display: grid;721 grid-template-columns: minmax(0, 0.75fr) minmax(0, 1.25fr);722 gap: 40px;723 align-items: start;724 }725 726 .thesis-quote {727 padding: 30px;728 border-left: 5px solid var(--spectrum-1);729 background: #fff;730 font-size: 25px;731 font-weight: 820;732 line-height: 1.32;733 letter-spacing: 0;734 }735 736 .capability-flow {737 display: grid;738 grid-template-columns: repeat(4, minmax(0, 1fr));739 gap: 10px;740 }741 742 .flow-step {743 position: relative;744 min-height: 160px;745 padding: 19px;746 border: 1px solid var(--line);747 border-radius: 3px;748 background: #fff;749 }750 751 .flow-step:not(:last-child)::after {752 position: absolute;753 top: 50%;754 right: -11px;755 z-index: 2;756 width: 20px;757 height: 20px;758 border: 1px solid var(--line);759 border-radius: 50%;760 background: #fff;761 color: var(--muted);762 content: ">";763 font-size: 12px;764 font-weight: 900;765 line-height: 18px;766 text-align: center;767 transform: translateY(-50%);768 }769 770 .flow-step b {771 display: block;772 margin-bottom: 14px;773 color: var(--blue);774 font-size: 12px;775 letter-spacing: 0;776 text-transform: uppercase;777 }778 779 .flow-step strong {780 display: block;781 margin-bottom: 7px;782 font-size: 17px;783 }784 785 .flow-step p {786 margin: 0;787 color: var(--muted);788 font-size: 13px;789 line-height: 1.5;790 }791 792 .track-grid {793 display: grid;794 grid-template-columns: repeat(2, minmax(0, 1fr));795 gap: 20px;796 }797 798 .scope-spectrum {799 position: relative;800 display: none;801 grid-template-columns: repeat(8, minmax(0, 1fr));802 gap: 8px;803 margin: -4px 0 24px;804 padding-top: 12px;805 }806 807 .scope-spectrum::before {808 position: absolute;809 top: 0;810 right: 0;811 left: 0;812 height: 4px;813 border-radius: 0;814 background: var(--line);815 content: "";816 }817 818 .scope-spectrum span {819 padding: 9px 7px;820 border: 1px solid var(--line);821 border-radius: 3px;822 color: #536078;823 background: rgba(255, 255, 255, 0.82);824 font-size: 11px;825 font-weight: 760;826 line-height: 1.25;827 text-align: center;828 }829 830 .track-card {831 position: relative;832 overflow: hidden;833 padding: 32px;834 border: 1px solid var(--line);835 border-radius: 4px;836 background: #fff;837 box-shadow: none;838 }839 840 .track-card::after {841 position: absolute;842 top: -65px;843 right: -65px;844 width: 190px;845 height: 190px;846 border-radius: 50%;847 content: "";848 display: none;849 opacity: 0.62;850 }851 852 .track-card.theory::after {853 background: #fff;854 }855 856 .track-card.practice::after {857 background: #fff;858 }859 860 .track-id {861 position: relative;862 z-index: 1;863 display: inline-flex;864 margin-bottom: 26px;865 padding: 7px 10px;866 border-radius: 3px;867 font-size: 11px;868 font-weight: 850;869 letter-spacing: 0;870 text-transform: uppercase;871 }872 873 .theory .track-id {874 color: var(--ink);875 background: var(--soft);876 }877 878 .practice .track-id {879 color: var(--ink);880 background: var(--soft);881 }882 883 .track-card h3 {884 position: relative;885 z-index: 1;886 margin-bottom: 6px;887 font-size: 30px;888 letter-spacing: 0;889 }890 891 .track-card .track-name {892 margin-bottom: 17px;893 color: var(--muted);894 font-weight: 720;895 }896 897 .track-card > p {898 max-width: 580px;899 color: var(--muted);900 }901 902 .track-metrics {903 display: grid;904 grid-template-columns: repeat(auto-fit, minmax(140px, 1fr));905 gap: 9px;906 margin: 24px 0;907 }908 909 .track-metric {910 padding: 14px;911 border-radius: 3px;912 background: var(--soft);913 }914 915 .track-metric strong {916 display: block;917 margin-bottom: 2px;918 font-size: 20px;919 }920 921 .track-metric span {922 color: var(--muted);923 font-size: 11px;924 font-weight: 650;925 }926 927 .tag {928 padding: 6px 9px;929 border: 1px solid var(--line);930 border-radius: 3px;931 color: var(--muted);932 background: #fff;933 font-size: 11px;934 font-weight: 720;935 }936 937 .track-link {938 display: inline-flex;939 margin-top: 24px;940 align-items: center;941 gap: 7px;942 font-size: 14px;943 font-weight: 820;944 }945 946 .theory .track-link {947 color: var(--purple);948 }949 950 .practice .track-link {951 color: var(--teal);952 }953 954 .figure-grid {955 display: grid;956 grid-template-columns: repeat(2, minmax(0, 1fr));957 gap: 18px;958 margin-top: 22px;959 }960 961 .figure-card {962 display: flex;963 overflow: hidden;964 flex-direction: column;965 border: 1px solid var(--line);966 border-radius: 4px;967 background: #fff;968 box-shadow: none;969 }970 971 .figure-card.wide {972 grid-column: 1 / -1;973 }974 975 .figure-media {976 width: 100%;977 min-height: 270px;978 max-height: 760px;979 object-fit: contain;980 background: #fff;981 }982 983 .result-figure {984 width: 100%;985 max-width: none;986 margin-left: 0;987 transform: none;988 }989 990 .result-figure .figure-media {991 max-height: none;992 }993 994 .figure-media.compact {995 min-height: 260px;996 max-height: none;997 }998 999 .caption {1000 flex: 1;1001 padding: 15px 18px 17px;1002 border-top: 1px solid var(--line);1003 color: var(--muted);1004 font-size: 13px;1005 }1006 1007 .caption {1008 display: none;1009 }1010 1011 .caption strong {1012 display: block;1013 margin-bottom: 3px;1014 color: var(--ink);1015 }1016 1017 .coverage-grid > .figure-card:not(.wide) .figure-media {1018 height: auto;1019 min-height: 320px;1020 object-fit: contain;1021 }1022 1023 .coverage-grid {1024 grid-template-columns: 1fr;1025 }1026 1027 .full-width-figure .figure-media {1028 display: block;1029 height: auto;1030 min-height: 0;1031 max-height: none;1032 object-fit: contain;1033 }1034 1035 .leaderboard-shell {1036 overflow: hidden;1037 border: 1px solid var(--line);1038 border-radius: 4px;1039 background: #fff;1040 box-shadow: none;1041 }1042 1043 .leaderboard-top {1044 display: flex;1045 gap: 18px;1046 align-items: center;1047 justify-content: space-between;1048 padding: 20px;1049 border-bottom: 1px solid var(--line);1050 }1051 1052 .leaderboard-date {1053 color: var(--muted);1054 font-size: 12px;1055 font-weight: 760;1056 white-space: nowrap;1057 }1058 1059 .sort-button {1060 display: inline-flex;1061 gap: 7px;1062 align-items: center;1063 justify-content: center;1064 width: 100%;1065 padding: 0;1066 border: 0;1067 color: inherit;1068 background: transparent;1069 cursor: pointer;1070 font: inherit;1071 font-weight: inherit;1072 line-height: 1.3;1073 text-transform: inherit;1074 }1075 1076 .sort-button:hover {1077 color: var(--blue);1078 }1079 1080 .sort-indicator {1081 color: #a4afbe;1082 font-size: 14px;1083 line-height: 1;1084 }1085 1086 .sort-button.active .sort-indicator {1087 color: var(--blue);1088 }1089 1090 .leaderboard-note {1091 max-width: 900px;1092 margin: 0;1093 color: var(--muted);1094 font-size: 14px;1095 line-height: 1.55;1096 }1097 1098 .tab {1099 padding: 9px 12px;1100 border: 1px solid var(--line);1101 border-radius: 3px;1102 background: #fff;1103 color: var(--muted);1104 cursor: pointer;1105 font-size: 13px;1106 font-weight: 780;1107 }1108 1109 .tab:hover {1110 color: var(--ink);1111 }1112 1113 .tab.active {1114 border-color: var(--navy);1115 background: var(--navy);1116 color: #fff;1117 }1118 1119 .table-wrap {1120 overflow-x: auto;1121 max-width: 100%;1122 -webkit-overflow-scrolling: touch;1123 scrollbar-gutter: stable;1124 }1125 1126 table {1127 width: 100%;1128 min-width: 720px;1129 border-collapse: collapse;1130 font-size: 13px;1131 }1132 1133 th,1134 td {1135 padding: 12px 14px;1136 border-bottom: 1px solid #edf0f4;1137 text-align: center;1138 white-space: nowrap;1139 }1140 1141 th {1142 color: #59667a;1143 background: #f8fafc;1144 font-size: 11px;1145 font-weight: 850;1146 letter-spacing: 0;1147 text-transform: uppercase;1148 }1149 1150 tbody tr:hover {1151 background: #fbfcfe;1152 }1153 1154 td.model,1155 th.model {1156 text-align: left;1157 }1158 1159 td.model {1160 color: var(--ink);1161 font-weight: 740;1162 }1163 1164 .rank {1165 color: var(--muted);1166 font-weight: 800;1167 }1168 1169 tbody tr:nth-child(1) .rank {1170 color: #a56b00;1171 }1172 1173 tbody tr:nth-child(1) {1174 background: #fffaf0;1175 }1176 1177 .score {1178 color: var(--blue);1179 font-weight: 850;1180 }1181 1182 .table-foot {1183 padding: 13px 18px;1184 color: var(--muted);1185 background: #fbfcfe;1186 font-size: 12px;1187 }1188 1189 .leaderboard-notes {1190 display: grid;1191 gap: 8px;1192 padding: 14px 18px 16px;1193 border-top: 1px solid var(--line);1194 color: var(--muted);1195 background: #fbfcfe;1196 font-size: 12px;1197 line-height: 1.55;1198 }1199 1200 .leaderboard-notes p {