Long-term ecological datasets are becoming increasingly abundant on the internet, and are available both on websites hosted by the originating sites and/or in online repositories. While sites and networks are increasingly conforming to adopted metadata standards (which themselves continue to evolve to meet modern data collection challenges), the data themselves are often found in seemingly infinite combinations of 1) various formats, ranging from fixed width format to complex Microsoft Excel workbooks, 2) various levels of quality, from raw, unchecked data to flagged and corrected data values, and 3) various levels of time-step aggregations, from the raw data that include every data point measured, to data aggregated into annual means, totals or other estimations. In addition, datasets are documented to varying levels of completeness. Thus, the time-consuming process of preparing data from different sites (or even from different studies) for synthetic analyses and visualization often falls on the researchers interested in making the comparisons, with support of principal investigators and information managers associated with the project or site of origination. Over time, many researchers may have to independently aggregate the same data. With increasing demand for these synthetic analyses, coupled with decreasing availability of time and funding, it is important that these datasets be made not only more accessible, but in forms that allow comparisons and visualizations to be made quickly and in a single website location. To respond to this demand, the EcoTrends project (http://www.ecotrends.info) was initiated in 2004 to synthesize long-term datasets (> 10 years) from many US sites. Here, we discuss how data were obtained from partnering sites; how they were handled, from origination to derivations being made available on the EcoTrends web portal; and how the website component of this project meets the needs of ecologists. We also discuss lessons learned along the way and present recommendations for future projects.