Historical instrumental observations are a valuable source of quantitative information on past climate variability, especially when printed records can be transformed into machine-readable datasets. This study uses meteorological observations published in the '<em>Annalen des Physikalischen Centralobservatoriums</em>' in Saint Petersburg to assess their potential for Eurasian climate analysis. We digitized monthly temperature and pressure data from printed tables and converted them into a machine-readable dataset, addressing challenges related to station metadata, spatial coverage, multilingual headings, historical station names, decimal notation, missing entries, and optical character recognition errors. Using selected stations, we construct monthly surface air temperature and pressure anomaly maps for 1877 and 1879 relative to station-specific monthly reference means based on the other available years within 1874-1894. The year 1877 falls within the major 1877-1878 global climate anomaly, while 1879 provides a contrasting case. The station-based anomaly maps show coherent spatial structures across large parts of Eurasia. In 1877, late-year positive pressure anomalies over central and eastern Eurasia are accompanied in December by widespread cold anomalies. In 1879, the maps show stronger regional contrasts and a markedly different December pressure pattern. Our results demonstrate that the <em>Annalen</em> observations are valuable not only as historical records, but also as quantitative data for studying nineteenth-century atmospheric variability.