In this poster, we report preliminary results from a survey of academic librarians in the U.S. and Canada on their perception of data literacy for their work practice. Our survey was distributed in August 2022 to librarians from member libraries of the Association of Research Libraries, Canadian Association of Research Libraries, Oberlin Group, as well as through various LIS listservs. We received 338 valid responses. Our focus is on the impact of library membership, librarians’ educational backgrounds, and percent of data work on librarians’ perception of the importance of various areas of data literacy. Significant library membership differences occurred both in the percentage of their job involving data-related tasks (H(3) = 9.146, p = .027), with ARL librarians having the highest mean rank, and in respondents’ importance rating on research data principles (H(3) = 10.534, p = .015), with CARL librarians having the highest mean rank. Librarians who had a non-MLS degree rated their data proficiency as significantly higher than respondents with MLS and another degree and people with only an MLS degree (H(2) = 8.815, p = .012). Percentage of data-related work was positively correlated with self-rated data proficiency (rho(n = 253) = .641, p = .000), the level of data literacy needed (rho(n = 262) = .352, p = .000), and various importance ratings of data literacy areas including programming skills (rho(n = 251) = .268, p = .000) and data processing & visualization (rho(n = 253) = .185, p = .003). Further analysis is currently in progress.