file dr7-densities.txt

columns: objid, ra, dec, redshift, density, min_density, max_density 
                                      (these are log sigma values)

ref. Baldry et al. 2006 method updated to DR7 data. 

The method uses the projected distance to the 4th and 5th nearest neighbours
that are within 1000km/s. The density-defining population (DDP) are galaxies
brighter than M_r<-20 (h=0.7).  The density, log sigma, is 0.5*log(sigma_4) +
0.5*log(sigma_5), where sigma_N=N/(pi*d_N^2) and d_N is the distance to the Nth
nearest DDP neighbour. See paper, section 2.6, for details.

column 1: SDSS objid from photometric catalog.

column 5: density is the average log density derived from the 4th and 5th
nearest neighbours, and with and without considering photo-z of neighbours with
no spec-z. When near an edge, the area, pi*d_N^2, is modified to take account of
the reduced photometric coverage.

column 6: min_density is density using distances to spectroscopically-confirmed
neighbours only.

column 7: max_density is density using distances to possible neighbours in
consideration of photo-z or the distance to the nearest photometric edge if
nearer.

Set limit on max_density - density, or max_density - min_density for
reliability.

In summary, this method is a standard cylinder, sigma, calculation with careful
attention paid to missing spec-z's from the SDSS main galaxy sample and to
photometric edges. The sigmas, usings M_r<-20 as the DDP, are "reliable" in the
range 0.01<z<0.085. Notes: (i) no attempt has been made to correct the densities
for missing photometric data within the SDSS legacy survey edges; there are a
few relatively large holes, and many holes around bright stars; (ii) there is no
correction for over-deblending, and missing large galaxies because of the
not-saturated criteria for example.