Include homeless populations.
In california it says we have 13+ million households. And that we have 14+ million housing units. That leaves us with over a million vacant units. Are the people who live in hotels, motels cars and rvs included. And are hotels motels and inn counted as part of the housing units? Because they should not be. Unless they are extended stays because they have no other place to rent. Not long term business trips. But used as actually housing. If there are over a million vacant units and we don't have that many households to fill them then why are whole families with kids on the streets? According to the labor board we only have 18 thousand jobs available but we have a population of 39 million. This does not add up at all. How can these numbers be so far off.
Unfortunately, the US Census Bureau does not generally include people experiencing homelessness in the American Community Survey.
For the decennial census, there are some efforts, described on this page https://www.census.gov/library/fact-sheets/2020/dec/2020-census-counts-homeless.html
The American Community Survey bases its estimate primarily upon people who receive a survey at their home mailing address. There are also efforts to count the "group quarters population" (see https://censusreporter.org/topics/group-quarters/ for some more) but that data is only released at the national level, and even there, the homeless population is not clearly separated from "other non-institutional".
I also found this 2012 book, "Small Populations, Large Effects: Improving the Measurement of the Group Quarters Population in the American Community Survey", which no doubt does much deeper (contents free online at https://www.nap.edu/read/13387/chapter/1 )
For the foreseeable future, Census Reporter is limited to data in the American Community Survey, but even were we to expand, it would likely only be to other official US Census sources.